ITQAN LAB

MIT · verified 2026-08-11

Updates

Every release of every tool, newest first. 87 releases across 15 tools.

An agent can read the same list as JSON at /updates.json, compare it with what is installed, and tell you exactly what changed before it updates anything.

dokploy-ops2.1.2

Changed

  • Code comments and setup examples use a placeholder project name (acme) instead of real ones.

hetzner-ops2.1.2

Changed

  • Code comments and setup examples use a placeholder project name (acme) instead of real ones.

namecheap-ops2.1.2

Changed

  • Code comments and setup examples use a placeholder project name (acme) instead of real ones.

plane-fleet3.0.0breaking

Breaking

  • scripts/lanes/run-lane.sh is removed. It only forwarded to the fleet command since 2.0.0. Call scripts/fleet (or core/scripts/fleet) instead.

Added

  • fleet schedule install --every 2h: unattended PM runs from a launchd agent (macOS) or a crontab line (Linux), with status, pause, resume and remove. run-once takes a lock, skips while paused or while an attended PM's heartbeat is recent, stops after fleet.unattended.max_hours and calls fleet.alert_command on failure. See core/references/unattended.md.
  • The unattended-guard hook, on only in an unattended run. It always denies a push, PR, merge or release tag on a production branch and a production deploy command. It allows merging and client messages only when fleet.unattended.allow lists them.
  • fleet schedule heartbeat, run by an attended PM each loop.
  • core/references/pm-history.md: the incidents and measurements behind the PM's rules.

Changed

  • An engineer run is scored only on a fail with cause missed. A fail the tester put on the ticket, the spec, the environment or a PM ruling is not scored either way.
  • selection.demote_window is 20 (was 5).
  • The PM runs fleet apply at preflight and after each merge group, and applies it.
  • preflight.md (25.5 to 17.8 KB) and spawn.md (27.3 to 19.1 KB) keep every rule, and their history moved to pm-history.md.

Fixed

  • spawn.md said both "do not log tokens" and "pass tokens" for a Claude spawn. It now says not to: tokens is the last call's size, and cost --backfill writes the real usage.

plane-fleet2.16.0

Added

  • fleet intake: every intake source's new items as JSON. A broken source shows ok: false and never fails the pass. See core/references/intake-sources.md.
  • A built-in Sentry source, on whenever the project config has a sentry block. It reads only the Sentry projects the block names, in its one environment. A new issue qualifies when it is first seen inside window_days or regressed, and passes min_users or min_events (any issue for always projects). Severity comes from users, urgent_paths and regressions. Evidence carries the link, counts, release, in-app frames and breadcrumbs, with query strings, emails, phone-like numbers and token=-style values removed.
  • Command sources in fleet.intake_sources, for any other feed. The command prints {"items": [...]}, and an optional resolve_command closes the item.
  • fleet intake check at run start. It warns when a repo ships a Sentry SDK but has no sentry block, so the PM tells the owner.
  • fleet intake mark and fleet intake resolve. A filed item is not offered again. A merged fix resolves the Sentry issue in the deployed release with a note naming the ticket, and an issue that comes back is offered again with regression_of set to the old ticket.
  • The token comes from $SENTRY_AUTH_TOKEN or a mode-600 token_file in the per-user override, never the repo.

Changed

  • PM preflight runs fleet intake check and the first pass. Each loop runs a pass before pre-pick triage. deploy-and-done resolves the source item. The planner and fleet-board have rules for intake items (external_source and external_id set, deduped on the board first).
  • On Plane MCP versions before the fix for PMCP-24, workitem list by external_source and external_id returns a validation error on a match instead of the row. The docs say how to read it.

plane-fleet2.15.0

Added

  • Rework cap in the spawn hook. An engineer spawn for a ticket with fleet.test_rounds_max failed test rounds (default 2, 0 turns it off) is refused until the owner's answer is recorded with fleet rounds allow <TICKET> --reason "...". One allowance covers one more round. fleet rounds lists the count per ticket. Switch file: rework-cap-off.
  • Two trials in fleet.trials, off by default: - case_list: before a build, a tester turns the ticket into a numbered CASE LIST: comment, with a state table for a screen and the grid each rule holds over. The engineer builds to it and reports pass or fail per case from a real run. A list over about 25 cases goes back to the planner to split. - closed_retest: a retest runs the failed cases and the list's regression, nothing new. A finding outside the list is graded s1, s2 or s3. Only s1 blocks (data loss, money, a wrong or duplicate send, security, a blocked main path). The rest become a follow-up ticket that lands before the release carrying the feature.
  • group.must_run and group.labels in the config. fleet gate-stamp check fails a group PR when a required job was skipped or missing, or a label is absent. A group PR had gone green in 17 minutes with its full suites skipped.
  • fleet slot verify-clean <S> and fleet slot release <S>. Engineers and testers paste the verify line as their teardown report. release never kills a process and refuses when a port or a database is in use.
  • fleet queue add | list | next: a spawn the machine gate refused is parked with its brief and offered again when there is room. "Room for N heavy" now also counts sessions already running.
  • fleet questions add | list | close: open owner decisions, the PR each one holds, and its age. Every status to the owner starts with them.
  • log add: --outcome spawn for the row written at the start of a session, so the result row needs no --force. --cause pm_ruling. --in-list, --severity and --repeat on a failure.
  • fleet status shows failed rounds per open PR, slots with leftovers, "required ran: x of y" on a group PR, queued spawns and open owner questions.
  • fleet retro counts PM-ruling failures (pm_ruling_failures), splits failures into new and raised before, reports how long queued spawns waited, and splits tester pass to merge into waiting for a group, CI and waiting for the owner.
  • core/pm/skills/ruling.md: a PM scope decision quotes the contract or spec line it rests on, or is marked a guess to verify, and lists the tests it retargets.
  • Brief checklist items 22 to 24: codes from an outside service, a fault in a shared component, two tickets in flight on the same locks.
  • Planner: one screen per UI ticket, a rule criterion names its grid, a contract change is an owner question in the ticket.
  • fleet.session_soft_cap (default 300) and fleet.client_engines.

Changed

  • The soft cap on self-review is 2 passes, down from 3, and comes after the cases are green.
  • An agent asked for progress answers with counts, never only "waiting".
  • fleet cost, fleet review and fleet retro do not count a spawn start row as a session.

plane-fleet2.14.0

Added

  • fleet held set|clear|list: the PM records each Todo ticket it skips for its blockers, from the relations it already read at pick time. No extra Plane call.
  • fleet status prints a held work line when every blocker of a held ticket has an open PR (draft, approved or in a group PR), a slot is free, and the machine is ok or tight with room. It names the ticket, the blocker PR and the branch to stack on. No line while a blocker has no pushed branch, no slot is free, or the machine is closed or critical. The compact-resume hook shows it too.
  • fleet retro counts the hours a ticket sat held with a slot free, from the sightings status writes: held_hours, a PM alert over half an hour. A run with nothing recorded says so instead of showing zero, since the outcome log has no field for it.
  • A STACKED ON block for the engineer brief: start from the blocker branch, rebase onto trunk after it merges, open the PR against trunk.

Changed

  • role.md and pick.md: a held-work line is answered in the same turn, by a stacked spawn or a reason logged as --outcome decision. Waiting for the blocker to merge is not a reason.
  • The engineer's worktree skill allows a blocker branch as the base when the brief has a STACKED ON block.

plane-fleet2.13.0

Added

  • fleet.trials in the config, all off by default. Turn one on only between runs. fleet review shows a "trials on" row, so each trial is judged against the runs without it. - path_map: on a money, stock, state or permission ticket the engineer posts a PATH MAP: comment before code, with every entry point and writer the rule touches. The blind tester may read it and checks every row in round 1. - money_ceiling: the planner splits money, stock and state outcomes at 1000 lines. - own_commits_only: a branch carries only its own ticket's work, in rework too. One ticket's branch took in a group branch and three other tickets' fixes, and went through 9 test rounds. - group_cut_minutes: the PM cuts a merge group no later than this many minutes after the first approval starts waiting. Approved to group opened was a median 1.5h, and for 54% of it no other group was open.
  • --defect on log add for a missed verify_fail: criterion, other_path, stale_state, concurrency, money_rule, out_of_scope or stacked. The tester names it in its report, and the review counts missed failures by kind.
  • fleet status warns about each ticket PR out of draft, with gh pr ready --undo. A draft runs almost no CI; ticket PRs out of draft used about 3,600 runner minutes in 18 days.

Changed

  • A verify_fail row needs --reason, like blocked. 67 of 186 failures had none.

plane-fleet2.12.0

Fixed

  • fleet cost and fleet retro left lane spend out. Lane tokens were read, but Codex had no price, and lanes were never added to Spend, so a Codex-only run showed agents at $0. Lanes are now priced and counted in Spend, the per-ticket cost and the PM share.
  • A lane row with no token counts was priced as $0. It is now counted as a session whose cost is unknown, and the report says how many there are.
  • fleet cost dropped the PM's last day: a window ending 2026-11-14 17:00 left out all of the 14th, while fleet retro kept it. PM spend now follows the window to the 10 minute, like the context per call already did. Whole days remain only for sessions known from an old cache entry.
  • fleet cost counted every PM session the repo ever had as the run's sessions (213 for a run that had 1). It now counts the sessions with calls in the window.
  • fleet retro showed no lost time both when nothing was lost and when fleet guard never ran. The guard now writes a start event, and the retro says "not checked" when no guard covered the window.

Added

  • Prices for gpt-6.1-sol and gpt-5.6-sol (OpenAI standard tier, short context, read 2026-11-16). A price can carry its own source and a valid_until date. fleet cost names a price whose date has passed. The gpt-5.6-sol price is promotional until 2026-11-30.
  • fleet review: finds the runs in the log, measures each with the same code as cost and retro, keeps one scorecard per finished run in {state_home}/reviews/<project>.jsonl, and prints the last runs side by side with what got better, what got worse, the toolkit versions released between runs, the session mix and the decisions logged. --md writes it for the owner's records. Config: fleet.review.
  • The review judges a run on three goals that count the same: time (merged tickets per run hour, spawn to merge), results (rework rate, merged of started, escapes logged) and cost ($ per merged ticket). A run better on one goal and worse on another is called a trade, never an improvement.
  • The review splits each merged ticket's time into stages: building, testing, rework, waiting for an engineer, a tester or rework, pass to merge, and waiting while no fleet session ran. It also gives cycle time and cost by rework rounds (0, 1, 2+).
  • Ticket timelines in the review use all of a ticket's sessions, and a ticket counts in the run where its first engineer or tester work started. A reviewer that only names the ticket, or a group merge, is not its work. Before, a ticket worked across two runs counted in both, and one reviewer mention moved a ticket's start 13 days early.
  • Merges are read per run (up to 7 days after it), not in one call capped at 1000 PRs. A run whose merges could not be read is not kept, and --rebuild keeps nothing when any run failed.
  • --outcome decision: a PM marker row for a routing choice or an owner's order, with a time and a reason. Never scored, never refused as a repeat.
  • The PM's retro skill has a review section: when it runs, what to do with it, and when to propose a bigger change than a tweak.

Changed

  • fleet log add refuses a blocked, lane_failed, deferred or decision row without --reason. A lane row takes the lane's last message when no reason is given, marked lane said:.

video-watch3.0.1

Fixed

  • On a Mac, a link failed with "No module named expat" before anything downloaded. Homebrew's Python can be built against a newer libexpat than the one macOS ships, so yt-dlp could not start. The script now checks yt-dlp's Python first, and when the module fails to load it uses the libexpat from Homebrew's expat formula. If that formula is missing, it says to run brew install expat.

zernio-ops1.0.0

Added

  • One MCP server, zernio, in front of every Zernio workspace. It exposes Zernio's tools once, adds an optional workspace argument to each, and finds the workspace from account_id when it is left out. accounts_list with no workspace lists them all.
  • The workspaces are the ZERNIO_KEY_<NAME> lines in the toolkit's credential store. There is no fixed count, and a name can be a letter or a word. A new workspace is one new line and a restart.
  • scripts/setup.mjs adds a key line for the person to fill in, checks each key against Zernio without printing it, and records the working workspaces in state/setup.json.

plane-fleet2.11.1

Fixed

  • fleet retro read the PM context per call for the whole UTC day, whatever --since said. It now follows the window, to the 10 minute. The cost figures stay counted by whole UTC day, as before. A session that only an old cache entry remembers still counts by day.
  • fleet findings reported an empty or missing review file as clean. It now says the review did not finish and exits 4, the same as a usage limit. A file with a marker but no final message does the same.
  • Waiting on a group PR with gh pr checks --watch could end early. A workflow can leave a skipped duplicate under the same check name, and --watch reads it as done while the real check still runs. The merge queue now waits on the run for the head commit (gh run watch <id> --exit-status) and confirms the required check concluded success on that head. fleet status and fleet gate-stamp check no longer let a skipped duplicate hide a running or passed check of the same name.
  • The fleet-planner and fleet-scanner agents could not read ticket comments. They now have workitem_comment, and their roles say they only list and retrieve.

Added

  • engineering/references/brief-checklist.md item 21: a rule about whether something may happen, or who may see or do something, gets one function every caller goes through, with the callers listed by grep before coding. It is numbered last so the numbers in existing briefs stay the same.

plane-fleet2.11.0

Added

  • fleet guard: one background process per run that holds off system sleep, checks that an agent call still works, and alerts through fleet.alert_command and a desktop notification.
  • fleet retro reports lost time (asleep, agents down, tester pass to merge) and a lost_hours finding.
  • ci.group_is_the_gate and quality.pre_push_full config: the group PR's CI is the one full run, and the hook can check only what the branch changed.

Changed

  • A run lands in a few bigger groups, and a chain of dependent tickets lands as one group.
  • blocked_by needs a named code dependency. A real chain is built stacked on the previous branch and does not wait for its merge.
  • The PM's focused review runs before the tester starts, not alongside it.

plane-fleet2.10.0

Added

  • engineering/references/brief-checklist.md: the defect classes that caused repeat rework, grouped (before code, UI, money, dates and concurrency, review before handoff, testers, PRs and CI). The PM copies the items that apply into each engineer brief under CHECKLIST:, and the engineer names in the handoff which ones it checked.
  • fleet.max_heavy_sessions (default 3, read by the PM): at most that many engineers or testers with servers and a test DB at once, even when the machine gate shows room. The gate's room figure ignores swap that is already high.
  • A gate lock: one full pre-push gate at a time on the machine. Sessions and the PM's group gate take /tmp/<identifier>-gate.lock with mkdir before a push and wait while it is held.
  • A release PR section in the merge queue: regenerate visual baselines on the CI runner when a release changes something drawn on every page, and never merge a [skip ci] commit last before a release PR.

Changed

  • Test workers are capped per session: half the cores divided by fleet.max_heavy_sessions, at least 1, in place of a fixed 2.
  • Ticket PRs stay drafts. Nobody marks one ready or adds the e2e label to it. CI runs on the group PR, and the full e2e suite runs on the release PR. The standing brief's old "add the e2e label before your last push" rule is gone.
  • The brief states per role who sees which screen, button and field, before round 1.
  • A PR past about 40 files, or with several features, is split before its next round.
  • Group PRs are checked for mergeable before anyone waits on them. A conflicting group gets the trunk merged in, the full gate again, and a look for clashes a text merge cannot see. A watcher with no checks after N minutes reports a fault.
  • Gate-stamp proof: a plain git pre-push hook does not print the line fleet gate-stamp post looks for. Until the script changes, the PM proves the gate by the remote branch head matching the local HEAD after a push without --no-verify. Nobody writes the log line by hand.
  • Stop-the-line names the runner first: PR checks can run on another runner than trunk jobs, a checkout error means a damaged work folder there, and pgrep inside sh -c matches itself.
  • An engineer whose ticket changes what a screen shows greps every e2e spec for the old text or figure and updates them in the same PR.

plane-fleet2.9.1

Fixed

  • fleet gate-stamp check read a stale re-run. gh pr checks keeps one row per check name. After a PR title edit it kept a re-run attempt of the old workflow run, which still carried the old title and failed, over the newer run that passed. The check said NOT READY while GitHub showed the PR as clean. It now also reads the PR's statusCheckRollup and, for each check name, uses the run with the highest workflow run id. An older passing run never hides a newer failure.

plane-fleet2.9.0

Added

  • fleet slot-template build|clone|status: one migrated, seeded Postgres template per schema state. build makes <prefix>_<key12> from origin/<trunk> in a throwaway worktree with the configured commands. The key hashes the migration and seed files and the build commands, not the commit, so a template is reused until they change. A lock stops two builds at once, and only keep templates stay. clone <slot> [--test] drops the slot's DB (and its test DB) and copies the template in seconds. It touches only slot DBs named in fleet.slots, never the shared row's, and exits 3 when no ready template matches so the session runs its recipe. Off unless fleet.slot_db_template.enabled.
  • fleet gate-stamp post|check: post puts a local-gate commit status on a SHA, only when HEAD is that SHA, the tracked tree is clean and the gate log shows success for it. check --pr N exits 0 when the PR head has the stamp and every check outside ci.covered_by_local_gate passed or skipped. A skip after a covered job was cancelled does not count as passed. Off while that list is empty.

Changed

  • Engineers and testers clone their slot DB with fleet slot-template clone and skip migrate and seed when it works. Otherwise they run the recipe.
  • Before any migrate, seed, reset or e2e command, echo DATABASE_URL and every test URL, check each names your slot DB, and export each in its own statement. On one project a combined export left the test URL on its old value, and the test harness migrated the owner's shared DB.
  • The PM rebuilds the template in the background after a group that changed migrations or seed.
  • A group PR into the trunk may merge when fleet gate-stamp check passes, even while a covered job still runs. Never the release PR to production, and never when the head moved after the stamp. Trunk CI still runs everything and stays the deploy gate.

plane-fleet2.8.0

Added

  • fleet cache-server start|stop|status|env: a small Nx remote cache on 127.0.0.1 (Nx 20.8+ self-hosted cache API), one per project, so every engineer, tester and group worktree reuses the others' lint, typecheck, unit-test and build results. Off unless fleet.nx_cache_server.enabled. eval "$(fleet cache-server env --cwd <repo>)" starts it and exports the two Nx variables. Oldest entries go past max_gb (default 20). Checked with two worktrees: the second got [remote cache].

Changed

  • Push once per session, after the last clean review pass. Review fixes stay local commits. Every push runs the whole pre-push gate and starts PR CI.
  • While working, run only the spec files you touched, never a whole project or suite, and nothing by hand right before a push.
  • Never run the same tests twice on one commit. While fixing a failure, re-run only the failing spec; run the full needed set once, after the last change. A rework or tester round re-runs only what its new commits touch.
  • Reports carry the push count and the number of full server or browser spec-set runs.
  • Testers push their test commits once, together.
  • Sharing one Nx cache folder across worktrees does not help on Nx 22 (each worktree's database ignores entries it did not write), which is why the cache server exists.
  • --skipNxCache is for a result that looks wrong for the change, not for every green: cached lint, unit tests and build are real results.
  • If the cache server is down, Nx runs the task and still exits 1 (Failed to send request ... /v1/cache). The brief says to restart the server and run again, never to skip the hook.

plane-fleet2.7.0

Added

  • fleet retro [--since ISO | --hours N] [--cwd repo] [--json]: per ticket rounds, spend, tool calls, time and outcome, totals, the PM's share and context per call, rework and its causes, lanes and the Codex burn rate, then numbered findings, each with a cause, a fix and who owns it (pm, toolkit, owner). Exit 3 on an alert. A finding seen in the previous window is marked REPEAT. Thresholds and levels are set under fleet.retro in config.
  • The PM's retro skill: one watcher runs the retro every 30 minutes, and again after each group merge and at the end of a run. The PM says each finding to the owner in one line, applies pm fixes at once and logs them as retro_fix, and proposes the rest with evidence.
  • fleet findings <file>: the final findings block of a codex review or codex exec transcript, deduplicated, with one count line. Exit 3 on an open P1 or P2, 4 when the review did not finish.
  • fleet watch --stop-prefix <prefix> stops every live watcher whose name starts with it.

Changed

  • The PM runs with auto-compact near 150k. Its state lives in Plane, the resume note and fleet status.
  • Every review transcript (the engineer's passes, the PM's focused review, the group review) goes to a file and is read only through fleet findings.
  • Engineers stop at a soft cap, 3 self-review passes or about 300 tool calls, and report CAP with what is done, what is open and the findings left. Testers stop at about 300 tool calls. The PM decides the next step.
  • Engineers and testers stop every background job they started before the final report. Watch names start with the ticket id.
  • An out-of-scope P1 is never pushed silently: the engineer comments the P1 and its scope argument and reports In Review · needs PM scope call. The PM decides before a tester starts.
  • Tickets with screenshots, visual baselines or e2e checks on counts and dates must say how each shot gets fresh data, a test clock and no leftovers, and ask for N unchanged reruns green.
  • Pre-pick triage: a missing cited path, or a named file, setting, preset or config that does not exist, sends the ticket to the planner to fix the text before anyone claims it. The planner checks each named thing exists before filing.
  • Merge groups stay small (about 5) when they carry a shared UI or dependency change, since per-PR CI skips most jobs on drafts.

Fixed

  • fleet log add refuses a row that repeats one already in the log (same agent id and outcome, or same project, ticket, role, executor and outcome within 10 minutes). --force logs a real second session.
  • Backfilled usage rows keep the failure cause and task kind of the row they copy, so a test failure no longer looks like a second failure without a cause.

plane-fleet2.6.0

Added

  • fleet watch <name> --until '<test>' and fleet watch <name> -- <command> run a background wait under a name. A second start under a name that is still running starts nothing and exits 75, with the holder's pid and output file. --show, --list and --stop read and end them. Engineers and testers wait only through it (standing brief), and the PM waits for CI with fleet watch <TICKET>-ci -- gh run watch <id>.
  • The sweep flags duplicate waits: the same until or while ... sleep loop running more than once from the repo. --kill keeps the newest copy.

Fixed

  • On macOS the gate no longer reads "swap nearly full" as critical. macOS adds 1 GB swap files as it needs them, so less than 1 GB left is normal once any swap is in use, and pages from apps open for days sit there at no cost. Swap in use now counts as tight at most. The kernel's memory-pressure level (warn closes, critical is critical) and the swap-out rate since the last probe (fleet.resources.swapout_mb_s, default 5, 25 and 100 MB/s) decide. Linux, Windows and a Mac that gives no pressure reading keep the old rule.

plane-fleet2.5.0

Added

  • fleet.tool_caches config: a short tmpdir for Jest, Nx and temp files, so a repo on an external disk stops filling the internal one. Lanes and gates get TMPDIR from it, fleet status names a Jest cache over its limit, and fleet-sweep.sh --caches (or the launchd job from fleet-cache-sweep.plist.template) clears old caches, temp folders and Docker's build cache. A Jest cache had reached 26 GB.

Changed

  • High-risk PRs (money, schema, API, auth, privacy, audit): a tester APPROVED does not count until the PM's focused review is back clean. That review runs on a different executor from the engineer's. The PM reads the engineer's open notes first, and any note touching totals, balances, permissions or written data is a must-fix before the tester starts.
  • Rework briefs ask for the whole class: every path, state and call site with the same shape, listed in the handoff as class checked:. Testers name the class with each finding, and the retest tries each one.
  • A rework that misses the same class twice moves to a different, stronger executor.
  • Two reworks is the cap, and hitting it counts as a fleet defect. Before a third round, the PM finds which step let it through (ticket, build, test, review or the PM's own briefs), logs it with a change to that step, and puts round 3 to the owner. Round 3 fixes the design, on a stronger executor.
  • A tester finding the ticket never asked for is a PM scope decision with a rule for the whole class, never a one-case Rework.
  • Rework and retest briefs point at one written case list on the ticket.
  • When group PRs skip the full suites, the group builder runs every test file that covers changed code, not only the specs the tickets name.
  • No release waits on one ticket by default. The PM tells the owner what holding costs, and a ticket with no user-visible change doesn't hold a release unless the owner says so.

plane-fleet2.4.1

Fixed

  • 2.4.0 wrote the worktree folders into an untracked .watchmanconfig ignore_dirs at the repo root. watchman then treats the repo root as the project for every worktree, so Jest inside a worktree found no tests and a pre-push hook passed green with three failing tests. fleet status and run-lane no longer write any watcher config. The warning now says to drop the watch and turn watchman off for Jest, and the PM docs forbid the ignore_dirs fix. If 2.4.0 wrote the file in a repo, delete <repo>/.watchmanconfig and the /.watchmanconfig line in .git/info/exclude.

plane-fleet2.4.0

Added

  • fleet resources (core/scripts/fleetres.py, standard library only). It measures memory the OS can still hand out, swap against RAM, load per core (CPU busy on Windows) and free disk on macOS, Linux (cgroup limits included) and Windows. Limits scale to the machine and can be set in fleet.resources. Levels: ok, tight, closed, critical (exit 0 to 3). It names the biggest processes, marks the fleet's own, and says what each usually is and whether its fix is safe (watchman, Nx daemons, headless browsers, VMs, browser tabs, indexers). It warns when a file watcher is rooted over the worktrees.
  • "Room for N heavy": how many more engineers or testers with a server, Jest or a browser the machine can take, from free memory less a reserve, capped by cores and max_heavy.
  • fleet resources --wait: blocks until the gate opens, at most 4 minutes. Briefs use it before Jest, Playwright or a server, in place of uptime arithmetic.
  • --gate heavy|light for scripts.
  • Briefs run Jest with --no-watchman.

Changed

  • run-lane defers a lane while the gate won't admit it (exit 75, resources <level> in the error). --ignore-resources skips the check.
  • The usage-pacer hook holds an engineer or tester Claude spawn on the same rule, and any fleet spawn at critical. The owner can switch it off with resources-off in PLANE_FLEET_HOME.
  • fleet status opens with a machine: line, the reasons, safe fixes for machine hogs, and watcher warnings.
  • fleet-sweep.sh prints the machine gate instead of the macOS-only check. --min-free-mem-pct and --min-disk-gb still override it for one run.
  • PM: a new hard rule, "never spawn past the machine gate", and a critical-level procedure: sweep, apply the safe fixes, tell the owner what is left, and only then stop the newest session's test run or server.
  • Standing brief: "Share the machine" uses fleet resources --wait, and no session starts a file watcher in a worktree.

plane-fleet2.3.0

Changed

  • The engineer's self-review now loops on P2s as well as P1s. When the brief carries the PM's focused review prompt, the engineer runs that one too, until one full pass reports neither. A P2 the engineer thinks doesn't apply goes to the PM before the push; on money, privacy, auth and audit tickets it blocks (engineering/engineer/skills/open-pr.md §3b).
  • For every finding, the engineer fixes the whole class: every writer of the field, every booking path, every caller, found by searching the whole repo. The sites checked go in the handoff, and the report line shows the focused pass count (engineering/engineer/role.md §6).
  • The PM writes the focused review prompt into every engineer brief: the ticket's invariants, the findings deferred to other tickets, and the failure kinds that matter (core/pm/skills/spawn.md).
  • The PM runs that review on every money, schema, API, privacy, auth and audit PR while the tester works, sends any finding to the running tester, and routes one Rework with everything instead of two (core/pm/skills/spawn-tester.md §4). On one project it found a P1 or P2 the tester's blind pass had missed on four PRs in a row.
  • After a third round, a ticket isn't patched again. A fresh opus engineer rebuilds it against three to five written invariants, or the planner splits it (core/pm/skills/pick.md §2).
  • A timed-out lane's commits are judged by the fallback engineer (keep, squash or rewrite) before anything is built on them (core/pm/skills/spawn.md §3, step 5).
  • A failed group PR fixed in place: easy fixes push first, each investigation gets its own lane, every lane gets a check at 30 minutes, and at most one other heavy session runs meanwhile (core/pm/skills/merge-queue.md §5).
  • The PM never sums up open findings as "nothing serious". Each one is named, with where it went (core/pm/role.md §7).

Added

  • A CPU load gate. No heavy spawn while the 1-minute load is over twice the core count, and at most 4 heavy sessions at once. Every session runs Jest at --maxWorkers=2 and waits out a high load before starting Jest, Playwright or a server (core/pm/skills/spawn-tester.md, engineering/references/standing-brief.md). Four heavy lanes plus two pre-push hooks had taken a 12-core Mac to a load of 108.
  • Ticket section §3c, invariants and every path they cover, required on money, concurrency and audit tickets. Rules the owner has to decide get decided there, not in review (core/planner/skills/write-ticket.md §3).

plane-fleet2.2.0

Changed

  • Schema tickets build in parallel when the models they touch don't overlap. Before, one schema ticket held every other schema ticket from build to deploy, about 1.5 to 2 hours each. Only the merge is still serial: one schema PR per train, and the next one rebases and re-stamps its migration folder before it merges (core/pm/skills/pick.md §2, merge-queue.md §5).
  • Engineer briefs on a schema ticket name the other schema tickets' models as off-limits and ask for an additive-only migration in its own folder. A migration that drops, renames or rewrites data still builds alone (engineering/engineer/skills/investigate.md).
  • Testers take the client screenshot and attach it through the Plane MCP only. When the MCP refuses a local file, they save it to ~/.cache/plane-fleet/shots/<IDENTIFIER>-<n>.png and name that path in the verdict comment. They never open Plane in a browser, never guess its web address, and never open a tunnel or upload host to get a public URL (engineering/tester/skills/verdict.md, engineering/tester/role.md, core/references/non-claude-adapter.md).

Added

  • fleet.parallel_schema in the project config: how many schema tickets may build at once. Default 2.

plane-fleet2.1.0

Added

  • Antigravity as a lane executor (fleet antigravity ..., the agy CLI on a Google AI plan). Models: gemini-4-argon, gemini-3.8-flash, claude-opus-5-5, claude-sonnet-5-5, gpt-oss-120b, with --effort. Opt-in: no candidates list names it by default.
  • A live meter for it: agy -p /usage, cached 5 minutes. Its three pools (Gemini, Claude and GPT, Argon) are judged apart, so pick and the lane gate look at the pool of the model being picked. fleet budget check antigravity --model <m> does the same by hand.
  • fleet antigravity-setup, run once: registers Plane with agy through fleet mcp-exec plane, which reads the server entry from ~/.claude.json at start so no key is copied, and links the lane guard plugin into agy.
  • The lane guard, an agy PreToolUse hook that acts only in fleet lanes. agy cannot set tools per run and its own denials end a headless run, so every lane runs with skip-permissions and the guard denies browser tools, other MCP servers and, in read-only lanes, writes, unlisted commands and Plane writes. A denial reaches the model and the lane still ends on its report. result.json says whether the guard ran and what it denied.
  • The adapter page covers Antigravity.

plane-fleet2.0.0breaking

Breaking

  • The plugin and its skill are renamed plane-fleet → fleet. Reinstall it under the new name. A linked install relinks skills/plane-fleet and each agent card.
  • The fleet is organized in sections: core/ (PM, planner, board specialist, scanner, solo mode, the fleet command, shared references), engineering/ (engineer, tester, their references and scripts) and social/ (planned). Every role is a folder with role.md and skills/, one file per thing it does, read only when that step comes.
  • The role skills team-pm, team-engineer, team-tester, team-ticket-writer and solo-ship are gone as separate skills. Start the PM with /plane-fleet run; the others are role files reached through fleet (table in SKILL.md). The ticket writer is now the planner's skills/write-ticket.md.
  • scripts/lanes/run-lane.sh is now the fleet command in core/scripts/, with scripts/fleet as its short path. The old path still forwards to it for this version.
  • Agent cards sit in agents/<section>/. fleet-plane is renamed fleet-board. A plugin install names them plane-fleet:<section>:<agent>.

Added

  • fleet-scanner, a read-only agent for counting, triage and log reading on behalf of the PM and planner.
  • A README per section.

Changed

  • Agent cards are short pointers; each role's procedure lives in its role.md.
  • The engineer's, tester's and PM's long role files are split into skill files by their existing sections, text unchanged. Section numbers (§3) stay, and references between files name the file.
  • The state folder keeps its name, ~/.config/plane-fleet (PLANE_FLEET_HOME), so the log, cost cache and overrides carry over. Hooks and cost still recognize sessions and log rows from before the rename.

plane-fleet1.6.0

Added

  • hooks/pm-recompact.py, a SessionStart hook on compact: in a PM session it adds the live run-lane.sh status and the PM's hard rules to the compacted context, so the PM continues from facts, not from the summary.

Changed

  • The PM runs with claude --autocompact 300k. Its context grew about 1.3k tokens a call, to 650k by call 400, and calls over 400k were 48% of PM spend; the default window on a 1M model compacts near 967k.

plane-fleet1.5.0

Added

  • hooks/cache-keeper.py: in fleet agents, refuses a foreground Bash timeout over 270000 ms and a Monitor deadline over 270000 ms (or none), and tells the agent to run the job in the background and check it every 4 minutes. An agent's prompt cache lives 5 minutes; waits past it made the next call rewrite the whole context, 12% of agent spend over a month and 24% in its last week. Refusals are logged; cache-keeper-off turns it off. cost shows the refusals next to the rewrite cost.
  • run-lane.sh status: budget, slots and who holds them, lanes running and just ended (a dead lane is flagged), open PRs with required checks, labels and mergeability, trunk CI, and worktrees with no open PR, in one call. The PM used about 2,000 separate calls a month for these.

Changed

  • The standing brief limits every wait to 4 minutes and says how to wait longer.
  • The PM never reads product code, runs builds or tests, or queries a database itself: 37% of its spend went there, on a context of 300k tokens and more. It reads the fleet with status and waits for CI with gh run watch in the background.
  • On a linked install, register cache-keeper.py in user settings like the pacer (README).

plane-fleet1.4.0

Added

  • run-lane.sh free-models refresh discovers the free models opencode serves today: no cost, tool calls, and at least min_context. It keeps each model's first and last day seen in {state_home}/free-models.json. run-lane.sh free-models lists them. New free models appear almost daily, some only for a short trial; 2 of the 6 listed on 2026-10-22 were gone three days later.
  • free_model_policy in the registry: the roles that may use free models (default scan; a free tier may keep the prompt), the providers to ask, and how old a list may be.

Changed

  • pick adds every live free model as a challenger for the allowed roles, and an experiment tries the newest untried one first. A model that is no longer served drops out. Every other role loses free models, whatever a list says.
  • run-lane refuses a free model outside the allowed roles or outside a read-only sandbox.
  • The registry's static free_models list is only a fallback when no discovery is fresh.
  • The PM runs free-models refresh in preflight.

plane-fleet1.3.0

Added

  • run-lane.sh cost: what tickets really cost and how long they took, read from the Claude Code session files of the PM and its agents, the lanes' usage and GitHub merge times. Per merged ticket: cost, agent + PM share, first spawn to merge, hours agents worked; cost by role and by rework rounds; where the money goes (cache reads and writes, cache rewrites after a wait, first-call context, the PM's calls and context). --ticket prints one ticket's sessions in order, --json gives the data.
  • cost --backfill [--dry-run] writes the real usage onto Claude spawn rows, so the scoreboard compares Claude and lanes on the same tokens.
  • A prices block in the executor registry: API list prices per model, used only by cost. A model without a price shows tokens only.
  • log add --cause missed|ticket|spec_changed|environment, required on every verify_fail. Of 114 test failures logged before this, none had a cause.
  • fleet.cost.claude_project_dirs in the project config, for a repo whose old sessions sit under another folder name.

Changed

  • The PM logs a Claude spawn's agent_id and duration, not its tokens, names every spawn <role> <TICKET>, and runs cost --backfill and cost in preflight.

Fixed

  • The tokens a Claude spawn row carried is the size of the agent's last call, not its spend: measured 2026-10-25, a median engineer session processed 9.8M tokens and a tester 14.3M, 44 and 56 times the logged figure. The docs that built cost claims on it are corrected: the agent types' start is about 17k tokens of context against 72k for general-purpose (1.2.0 said 23k against 40k), and the fixed cost of a ticket is now given in measured dollars.
  • cost keeps a summary of each session it reads, so a session still counts after Claude Code deletes its file (30 days by default).
  • A repo that moved had its old run-log rows counted twice: the copy under the new Claude projects folder got a new source_key. On one project that was 459 of 1,442 rows, so the scoreboard, reports and promotions worked on doubled samples, and seed would copy the rows in again. A legacy row's identity no longer includes the folder.

plane-fleet1.2.0

Added

  • Four agent types in agents/: fleet-planner plans right-sized tickets, fleet-plane files and reshapes them on Plane, fleet-engineer and fleet-tester run the existing roles. Each carries only the tools its role uses. Measured on one project: a spawn starts at about 23k tokens, against about 40k for general-purpose.
  • Shape rules in engineering/references/ticket-rules.md §2: one outcome per ticket, end to end; never split by layer while under the ceiling; siblings on the same screen or component ship together.
  • opencode lanes can use opencode's free models and the Z.ai coding plan as provider/model. Free models cost 0 in the cost units. Model families ignore the provider prefix.
  • disable_plugins in the executor registry turns off Codex plugins in a lane. The browser and computer-use plugins bring cua_repl, which drives the owner's real Chrome and desktop.
  • Every lane writes a small status file under {state_home}/lanes/, so a status view can list live and finished lanes without knowing each lane's --out.

Changed

  • The PM spawns fleet-engineer and fleet-tester by subagent_type, and sends new work through the planner and the Plane specialist instead of writing tickets itself.
  • The ticket floor is about 150 changed lines, up from 50.
  • The usage pacer also gates spawns of the four agent types.
  • fleet-sweep.sh finds slot scratch dirs by fleet.slot_scratch_prefix (default /tmp/<identifier>-slot-), not one hard-coded project name.
  • Examples in the rules and roles name no project.
  • Effort is part of a candidate: executor:model@effort is scored, explored and promoted on its own. pick returns it, log add --effort and lane results record it, and opencode lanes pass it as --variant.
  • Roles on another coding CLI never open Plane in a browser, never control the owner's desktop or browser, and never upload a file to an outside host to get a link. An attachment the MCP refuses stays local.

Fixed

  • opencode lanes logged no tokens: the runner now sums opencode's per-step token counts, so their cost and size show on the scoreboard.
  • A row logged with kind * was counted twice on the scoreboard.

plane-fleet1.1.0

Added

  • engineering/scripts/fleet-sweep.sh finds what finished or dead sessions left running (servers on free slots' ports, processes in removed worktrees, Nx daemons, old headless Chromium, wait loops that can never end) and stops it with --kill. It also prints a spawn gate on free memory and disk. The PM runs it before every spawn.
  • run-lane.sh report --project <p> --since <iso>: tokens per Done ticket, first-pass approval, rework rounds and the share of Done work the client can see, for one run.
  • run-lane.sh log add --from-result <result.json> takes a lane's tokens and duration from its own file. selftest checks the log round trip.
  • Rework rounds by ticket kind in engineering/references/ticket-rules.md.

Changed

  • Testers run targeted specs only. The PM's group PR runs the full server and browser suites, opened as a draft, labelled, then marked ready, so CI sees the label.
  • The standing brief asks every session to leave nothing running: one browser per session, closed in a finally, no Nx daemon, and a teardown that checks ports, processes and daemons.

Fixed

  • log add --set dropped its values on spawn rows, so the outcomes log had outcomes but no tokens or durations.

audio-narrate1.0.0

Added

  • Turn text into English or Arabic narration for free, with natural online voices, or offline with Piper.
  • Make one file per line of a script, named for the scene order, to cut a video to.
  • Fix words the voice says wrong with a pronunciation list, built in or kept in the project.
  • Set up once with a check that says what is missing and how to install it. The engines and voices live outside the skill folder, and the offline voices download only when asked for.
  • Clear messages when the free online voice service refuses, or when Piper's install path is too long, with what to do next.

audio-transcribe1.0.0

Added

  • Transcribe a local audio or video file, a direct media link, or a YouTube or other site link, fully on this machine.
  • Arabic (including Egyptian dialect), English, or automatic language detection.
  • Plain text, SRT and VTT subtitles, or JSON with timings.
  • Uses the fastest local Whisper engine installed: mlx-whisper on Apple Silicon, whisper-ctranslate2 or openai-whisper elsewhere.
  • A model is never downloaded without first saying how big it is.
  • Warns when the audio is near-silent, because Whisper invents text over silence.
  • Retries once with the large model when the output falls into a repetition loop.
  • A dependency check that shows the install commands and asks before running them, on macOS, Linux and Windows.

cloudflare-ops2.1.1

Fixed

  • Running lib/store.mjs through a symlinked install now works. Before, it exited without doing anything, because the check for being run directly compared the link with the real file.

dokploy-ops2.1.1

Fixed

  • Running lib/store.mjs through a symlinked install now works. Before, it exited without doing anything, because the check for being run directly compared the link with the real file.

elevenlabs-studio1.2.0

Added

  • A free check of a finished take: the median pitch against an expected range, and every word of the script heard back by a local transcriber. Exit code 5 means redo the line. Arabic is compared without diacritics, and --dialect eg also accepts Egyptian qaf spoken as a glottal stop.

Fixed

  • Running scripts/eleven.mjs through a symlinked install now works. Before, it exited without doing anything.

hetzner-ops2.1.1

Fixed

  • Running lib/store.mjs through a symlinked install now works. Before, it exited without doing anything, because the check for being run directly compared the link with the real file.

higgsfield-studio1.1.0

Added

  • A project can pin the Higgsfield account in .higgsfield/config.json (or HF_ACCOUNT). A paid job on any other signed-in account is refused before it runs, since the login is machine-wide and a free account looks just like the paid one.
  • The reserve can be set per project in the same file.
  • create no longer waits inside the CLI. It saves the job id at once and polls with a deadline (--timeout, 900 s by default). If it stops early, running the same command again picks the job up without paying again.

Fixed

  • Paid jobs run one at a time per project, so two jobs started together no longer count each other's charges.
  • A fresh sign-in with no billing workspace selected now says how to pick one.
  • Running scripts/hf.mjs through a symlinked install now works. Before, it exited without doing anything.

image-prompt1.0.1

Fixed

  • Running scripts/prompt.mjs through a symlinked install now works. Before, it exited without writing anything.

image-prompt1.0.0

Added

  • Write an image prompt in the project's brand, from a brand file the project owns. No brand file is fine.
  • Four directions: editorial, clean studio, detail and an overlay with room for a headline.
  • Exact sizes for product pages, Instagram, Facebook, LinkedIn, X, Pinterest and YouTube thumbnails.
  • A JSON beside each prompt with the size, ratio, don't list and a suggested Higgsfield model, with the crop needed when a model cannot make the exact ratio.
  • Short paste steps for any web image tool.
  • A prompt log per project that stops the same prompt being written twice.

namecheap-ops2.1.1

Fixed

  • Running lib/store.mjs through a symlinked install now works. Before, it exited without doing anything, because the check for being run directly compared the link with the real file.

toolkit-credentials2.1.1

Fixed

  • Running scripts/store.mjs through a symlinked install now works. Before, it exited without doing anything.

toolkit-updates2.2.0

Added

  • A tool installed under an old name is now reported as an update to its new name, with what to remove and what to install. Before, it was skipped without a word.

video-prompt1.0.1

Fixed

  • Running scripts/prompt.mjs through a symlinked install now works. Before, it exited without writing anything.

video-prompt1.0.0

Added

  • Write a short video prompt in the project's brand, from a brand file the project owns. No brand file is fine.
  • Five shots: turntable, hero reveal, editorial, detail, and animating an existing photo.
  • Sizes and accepted lengths for Reels, Stories, TikTok, YouTube, Shorts, LinkedIn, Facebook and product pages. A length a platform refuses is refused here too.
  • A JSON beside each prompt with size, ratio, length and a suggested Higgsfield model, a cheaper draft model, and how many clips a longer piece needs.
  • Short paste steps for any web video tool.
  • A prompt log per project that stops the same prompt being written twice.

video-watch3.0.0breaking

Breaking

  • Renamed from watch-video to video-watch, to follow the toolkit's naming rule: our own tools are named medium first, then action. Remove the old copy, then install video-watch: /plugin install video-watch@itqan in Claude Code, codex plugin add video-watch@itqan in Codex, or ./scripts/install.sh video-watch for any other agent. Nothing else changed.

elevenlabs-studio1.1.0

Added

  • Transcribe a recording or a link, with speaker labels, word timings and a list of names to spell right.
  • Change the voice in a recording and keep the performance. Can clean the noise in the same call.
  • Remove background noise and music from a recording.
  • Dub a video or audio file into another language, in voices cloned from the speakers. An interrupted dub resumes instead of paying twice.
  • Support for eleven_v4, ElevenLabs' newest voice model. Measured at about 0.11 credits per character, a quarter of eleven_v3.

Changed

  • Voice-over now uses eleven_v4 by default, as ElevenLabs does. Pass --model eleven_v3 for the old default.
  • Calls priced by the length of a recording refuse to run when the length is unknown. Install ffmpeg, or pass --seconds.

higgsfield-studio1.0.0

Added

  • Make images, video, audio and 3D with Higgsfield from the agent, through Higgsfield's own command line tool.
  • Every paid job is priced by Higgsfield before it runs, refused above --max or below the reserve, and never paid for twice.
  • Each job's real cost is read from Higgsfield's transaction list and written to a ledger in the project.
  • Product photoshoots and marketplace cards are planned and priced for free before any image is made.
  • Deploying or publishing a website, training a Soul and changing the billing workspace need the user's yes.
  • Seven more skills in the plugin, built on Higgsfield's official ones: Soul ID, product photoshoot, marketplace cards, brand kit, YouTube thumbnails, video explainers and websites.

plane-fleet1.0.0

Added

  • Five roles that work a Plane board together: a PM that claims tickets oldest first and runs the merge queue, engineers that each take one ticket to a draft PR in their own worktree and environment slot, testers that check a PR blind in the browser and then add tests, a ticket writer, and a solo mode for one session.
  • Project config in .plane/config.json, with a private per-user override for client contacts, server names and timezone, and run-lane.sh config to print the merged result.
  • Executors: any role can run as a sub-session or as a background lane on another coding CLI, picked per spawn by role, task kind and remaining budget. Lanes write a result.json with status, report line, duration and token use, and never commit on their own.
  • One budget pacer for every executor: a hard stop and a pace rule on each meter, pacer-off and pace-off switches, and a lane that has no budget exits as deferred with a resume time.
  • One outcomes log for all projects, a scoreboard per role, task kind and executor, and automatic promotion or demotion of executors, written to a decisions file the owner can veto.
  • A hook that holds new fleet sub-sessions when the usage window is ahead of pace.
  • A Plane MCP server declared by the plugin, run from source with uvx.

elevenlabs-studio1.0.0

Added

  • Connect an ElevenLabs account through a guided setup. The key is pasted into a file, never into the chat.
  • Make voice-over with any voice, with per-character timing for cutting a picture, and with neighbouring lines passed along so separate takes sound like one read (on models that support it).
  • Make sound effects at an exact length, including loops that repeat without a gap.
  • Make music from a prompt or from a composition plan. Plans are free, and the default model holds each section to its length.
  • List voices and save their preview samples for free, to choose a voice by ear.
  • See the plan, the credits left and the reset date.
  • Every paid call is estimated first, served from a cache when it was already paid for, measured after, and written to a ledger in the project. Voice-over cost is measured exactly from ElevenLabs' history. A reserve refuses any call that would leave too few credits.

cloudflare-ops2.1.0

Added

  • The tool has its own icon, in a dark and a light version. Codex shows it on the plugin card and in the skill list, and the toolkit site shows it on the tool's page and in share previews.

cloudflare-ops2.0.0breaking

Breaking

  • Installed as a Claude Code plugin, the slash command is now /cloudflare-ops:cloudflare-ops, with the plugin name in front. This has been true since 1.2.0, and this release marks it as breaking. Use the new name, or ask in words, which works as before. A copy in a skills folder keeps the plain name /cloudflare-ops.

Added

  • Codex now shows a display name, a short line and a starter prompt for the skill in its skill list. They come from a generated agents/openai.yaml inside the skill folder. Other agents ignore that file.

Changed

  • The description now leads with what the skill does, then its trigger phrases, then the detail. Codex shortens long skill lists from the end, so the phrases that route a request now survive far longer. With 40 skills installed, 95% of the phrases stay visible instead of 8%.

dokploy-ops2.1.0

Added

  • The tool has its own icon, in a dark and a light version. Codex shows it on the plugin card and in the skill list, and the toolkit site shows it on the tool's page and in share previews.

dokploy-ops2.0.0breaking

Breaking

  • Installed as a Claude Code plugin, the slash command is now /dokploy-ops:dokploy-ops, with the plugin name in front. This has been true since 1.2.0, and this release marks it as breaking. Use the new name, or ask in words, which works as before. A copy in a skills folder keeps the plain name /dokploy-ops.

Added

  • Codex now shows a display name, a short line and a starter prompt for the skill in its skill list. They come from a generated agents/openai.yaml inside the skill folder. Other agents ignore that file.

Changed

  • The description now leads with what the skill does, then its trigger phrases, then the detail. Codex shortens long skill lists from the end, so the phrases that route a request now survive far longer. With 40 skills installed, 95% of the phrases stay visible instead of 8%.

hetzner-ops2.1.0

Added

  • The tool has its own icon, in a dark and a light version. Codex shows it on the plugin card and in the skill list, and the toolkit site shows it on the tool's page and in share previews.

hetzner-ops2.0.0breaking

Breaking

  • Installed as a Claude Code plugin, the slash command is now /hetzner-ops:hetzner-ops, with the plugin name in front. This has been true since 1.2.0, and this release marks it as breaking. Use the new name, or ask in words, which works as before. A copy in a skills folder keeps the plain name /hetzner-ops.

Added

  • Codex now shows a display name, a short line and a starter prompt for the skill in its skill list. They come from a generated agents/openai.yaml inside the skill folder. Other agents ignore that file.

Changed

  • The description now leads with what the skill does, then its trigger phrases, then the detail. Codex shortens long skill lists from the end, so the phrases that route a request now survive far longer. With 40 skills installed, 95% of the phrases stay visible instead of 8%.

namecheap-ops2.1.0

Added

  • The tool has its own icon, in a dark and a light version. Codex shows it on the plugin card and in the skill list, and the toolkit site shows it on the tool's page and in share previews.

namecheap-ops2.0.0breaking

Breaking

  • Installed as a Claude Code plugin, the slash command is now /namecheap-ops:namecheap-ops, with the plugin name in front. This has been true since 1.2.0, and this release marks it as breaking. Use the new name, or ask in words, which works as before. A copy in a skills folder keeps the plain name /namecheap-ops.

Added

  • Codex now shows a display name, a short line and a starter prompt for the skill in its skill list. They come from a generated agents/openai.yaml inside the skill folder. Other agents ignore that file.

Changed

  • The description now leads with what the skill does, then its trigger phrases, then the detail. Codex shortens long skill lists from the end, so the phrases that route a request now survive far longer. With 40 skills installed, 95% of the phrases stay visible instead of 8%.

toolkit-credentials2.1.0

Added

  • The tool has its own icon, in a dark and a light version. Codex shows it on the plugin card and in the skill list, and the toolkit site shows it on the tool's page and in share previews.

toolkit-credentials2.0.0breaking

Breaking

  • Installed as a Claude Code plugin, the slash command is now /toolkit-credentials:toolkit-credentials, with the plugin name in front. This has been true since 1.2.0, and this release marks it as breaking. Use the new name, or ask in words, which works as before. A copy in a skills folder keeps the plain name /toolkit-credentials.

Added

  • Codex now shows a display name, a short line and a starter prompt for the skill in its skill list. They come from a generated agents/openai.yaml inside the skill folder. Other agents ignore that file.

Changed

  • The description now leads with what the skill does, then its trigger phrases, then the detail. Codex shortens long skill lists from the end, so the phrases that route a request now survive far longer. With 40 skills installed, 95% of the phrases stay visible instead of 8%.

toolkit-updates2.1.0

Added

  • The tool has its own icon, in a dark and a light version. Codex shows it on the plugin card and in the skill list, and the toolkit site shows it on the tool's page and in share previews.

toolkit-updates2.0.0breaking

Breaking

  • Installed as a Claude Code plugin, the slash command is now /toolkit-updates:toolkit-updates, with the plugin name in front. This has been true since 1.1.0, and this release marks it as breaking. Use the new name, or ask in words, which works as before. A copy in a skills folder keeps the plain name /toolkit-updates.

Added

  • Codex now shows a display name, a short line and a starter prompt for the skill in its skill list. They come from a generated agents/openai.yaml inside the skill folder. Other agents ignore that file.

Changed

  • The description now leads with what the skill does, then its trigger phrases, then the detail. Codex shortens long skill lists from the end, so the phrases that route a request now survive far longer. With 40 skills installed, 95% of the phrases stay visible instead of 8%.

Fixed

  • A skill header with Windows line endings (CRLF) is read correctly, so the version is no longer misread on a Windows checkout.

video-watch2.1.0

Added

  • The tool has its own icon, in a dark and a light version. Codex shows it on the plugin card and in the skill list, and the toolkit site shows it on the tool's page and in share previews.

video-watch2.0.0breaking

Breaking

  • Installed as a Claude Code plugin, the slash command is now /watch-video:watch-video, with the plugin name in front. This has been true since 1.3.0, and this release marks it as breaking. Use the new name, or ask in words, which works as before. A copy in a skills folder keeps the plain name /watch-video.

Added

  • Codex now shows a display name, a short line and a starter prompt for the skill in its skill list. They come from a generated agents/openai.yaml inside the skill folder. Other agents ignore that file.

Changed

  • The description now leads with what the skill does, then its trigger phrases, then the detail. Codex shortens long skill lists from the end, so the phrases that route a request now survive far longer. With 40 skills installed, 95% of the phrases stay visible instead of 8%.

cloudflare-ops1.2.0

Added

  • Can be installed as a Codex plugin, from the same source as everything else.

Changed

  • The changelog address now points to the site, so it keeps working if the repository layout changes.

cloudflare-ops1.1.0

Added

  • This changelog, so you can see what changed in each version.
  • An Updates section in SKILL.md. It tells the agent how to check for a newer version, and to report what changed before it updates.

dokploy-ops1.2.0

Added

  • Can be installed as a Codex plugin, from the same source as everything else.

Changed

  • The changelog address now points to the site, so it keeps working if the repository layout changes.

dokploy-ops1.1.0

Added

  • This changelog, so you can see what changed in each version.
  • An Updates section in SKILL.md. It tells the agent how to check for a newer version, and to report what changed before it updates.

hetzner-ops1.2.0

Added

  • Can be installed as a Codex plugin, from the same source as everything else.

Changed

  • The changelog address now points to the site, so it keeps working if the repository layout changes.

hetzner-ops1.1.0

Added

  • This changelog, so you can see what changed in each version.
  • An Updates section in SKILL.md. It tells the agent how to check for a newer version, and to report what changed before it updates.

namecheap-ops1.2.0

Added

  • Can be installed as a Codex plugin, from the same source as everything else.

Changed

  • The changelog address now points to the site, so it keeps working if the repository layout changes.

namecheap-ops1.1.0

Added

  • This changelog, so you can see what changed in each version.
  • An Updates section in SKILL.md. It tells the agent how to check for a newer version, and to report what changed before it updates.

namecheap-ops1.0.0

Added

  • Connect one or more Namecheap accounts. Each one is kept apart by name.
  • List domains, see what expires and what will not renew by itself, and check the balance.
  • Look up nameservers and change them.
  • Check whether a name is free, and see registration and renewal prices.
  • Edit DNS records for domains that use Namecheap's own DNS.
  • Handle Namecheap's address whitelist, including through an SSH server with a fixed address.

toolkit-credentials1.2.0

Added

  • Can be installed as a Codex plugin, from the same source as everything else.

Changed

  • The changelog address now points to the site, so it keeps working if the repository layout changes.

toolkit-credentials1.1.0

Added

  • This changelog, so you can see what changed in each version.
  • An Updates section in SKILL.md. It tells the agent how to check for a newer version, and to report what changed before it updates.

toolkit-updates1.1.0

Added

  • Finds copies installed as Codex plugins, and gives the Codex update command for them.
  • Finds Claude Code plugin copies in the newer layout, where the skill sits in a skills/ folder. The older layout still works.
  • Each line says where the copy lives, so a skill installed in two agents is easy to tell apart.
  • Can be installed as a Codex plugin, from the same source as everything else.

Changed

  • Claude Code keeps old versions in its plugin cache after an update. Only the newest one is checked now, so an old folder is no longer reported as behind.
  • The changelog address now points to the site, so it keeps working if the repository layout changes.

toolkit-updates1.0.0

Added

  • Check every toolkit skill installed on the machine against the public update feed, and list what changed in each newer version.
  • Finds copies in the folders agents read, in a project, in Claude Code's plugin cache, and behind symlinks.
  • Marks a breaking change, and gives the exact update command for each copy.
  • Read-only. Exit code 10 means an update is available, and 2 means the feed could not be read.
  • Optional --available lists toolkit skills that are not installed yet, and --json gives structured output.

video-watch1.3.0

Added

  • Can be installed as a Codex plugin, from the same source as everything else.

Changed

  • The changelog address now points to the site, so it keeps working if the repository layout changes.

video-watch1.2.0

Added

  • This changelog, so you can see what changed in each version.
  • An Updates section in SKILL.md. It tells the agent how to check for a newer version, and to report what changed before it updates.

dokploy-ops1.0.0

Added

  • Connect one or more self-hosted Dokploy installations. Each one is kept apart by name, so a deploy cannot land on the wrong server.
  • See every app, database and compose stack with its real state.
  • Deploy, and wait for the build to finish.
  • Read logs, restart and stop services.
  • Check domains and environment variables.
  • Adopt a key the machine already has.

hetzner-ops1.0.0

Added

  • Connect one or more Hetzner Cloud projects. Each one is kept apart by name.
  • List, create, start, stop, resize and delete servers.
  • Manage SSH keys, firewalls, volumes and private networks.
  • Manage DNS zones and records.
  • Show what the project is costing.
  • Adopt a token the machine already has.

cloudflare-ops1.0.0

Added

  • Connect a Cloudflare account in two steps. The token needs three checkboxes and is never typed into the chat.
  • Manage DNS records and subdomains.
  • Publish a Pages site from start to finish, with a custom domain.
  • Deploy a Worker and route a hostname to it.
  • Manage R2 buckets.
  • Look up any Cloudflare API endpoint instead of guessing at it.
  • Declare the Node requirement, and offer to install Node when it is missing.

toolkit-credentials1.0.0

Added

  • One shared place to store API keys and tokens, on macOS, Windows and Linux.
  • A guided walk through getting a credential from the provider, then a check that it works.
  • The secret is never typed into the conversation.
  • Declare the Node requirement, and offer to install Node when it is missing.

video-watch1.1.1

Changed

  • The docs now say which platforms it runs on: macOS and Linux, and Windows through WSL or Git Bash.

video-watch1.1.0

Added

  • A setup wizard that checks for ffmpeg and yt-dlp and installs them.
  • Automatic transcription when a source has no captions.
  • An optional free local engine for transcripts that runs offline.

Fixed

  • Playlists and audio-only sources are handled.

video-watch1.0.0

Added

  • Watch a video from a URL or a local file. It pulls captions and extracts frames the agent can read.
Itqan Lab

Built at Itqan Lab, a design and technology studio.

إتقان — itqan, the Arabic word for mastery: doing a thing precisely, and completely.

Open source under MIT · agent paths re-verified 2026-08-11 · this site is generated from the repository on every push.