Decide where work runs and whether this Mac can carry it yet. Routes across five execution planes — locally, an anvil errand container, a proctor instrument session, another model via defer:defer, or in-session subagents — by spending whichever resource is least scarce. Governs the local plane with berth admission…
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
End-to-end skill-improvement pipeline: source skill + user feedback + paid/free Dossier deep-research panel (read in full, citations verified) merged into a plan; the improved skill built as a new plugin; proven with comparative structural evals and a blind multi-family AI judge panel (CLIs + APIs, keys via 1Password).
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
Turn repository briefs, plans, code and mocks into traceable OVERVIEW.md and PRD.md documents, then an interactive launch site grounded in the actual product. Runs four phases: product synthesis, positioning and copy, implementation through design-craft:design-craft and ux-craft:ux-craft, and rendered acceptance. Uses.
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
Design and review authentically native macOS application interfaces — mock windows as self-contained HTML/CSS or a token-precise spec, with a committed aesthetic direction, a state matrix, a token table an implementer can build from, and a deterministic gate that computes contrast instead of asserting it. Grounded in…
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
Gets everything load-bearing out of a session and onto the page before the context holding it is thrown away. Measured across 121 real compaction events, the built-in prompt keeps 0.3% of the approaches you ruled out and 33.8% of your standing constraints; violations run 0% when a rule survives into the summary and…
★not rated 2
changed 21d agoA
tokens not measured
originalMIT
Incrementally digest macOS app UI screenshots, app icons, and official Apple UI kits into a persistent design corpus — per-app token profiles, cross-app pattern entries, aesthetic style clusters, icon anatomy digests, and a master TASTE.md — where every value carries a provenance mark from a closed set and every…
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
Tiered macOS maintenance for a dev machine that agent CLIs, node/dev servers, Docker and Xcode have quietly filled up. Five launchd cadences (15m/1h/12h/1d/7d) whose autonomy widens as the cadence lengthens: short tiers act silently on provably-regenerable debris, long tiers propose and wait. Verification gates before.
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
Validate that an implemented React or React Native UI faithfully reproduces a reference mockup, then update the code to close every gap, by measuring the rendered tree rather than eyeballing it or reading source. Treats the mock as the source of truth, inventories every frame including the modals and states the named…
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
Decide what a product should stand for, and be able to show your working. Runs its own Dossier deep-research panels (free CLI lane plus paid API lane, one or more panels decomposed by archetype) rather than emitting prompts for a human to paste elsewhere; generates candidate positions under trawl's persona frame…
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
Run a real test campaign against a native macOS app or an iOS Simulator: exploratory sweep, acceptance criteria turned into executable flows, state-matrix and edge coverage, accessibility audit, visual fidelity, determinism measurement, and a report that separates what was proven from what was assumed. Ships with the…
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
Work out what actually remains in a project by reconciling what it promised against what anybody proved — the feature briefs and PRD in docs/features-to-triage on one side, a test-campaign:test-campaign registry of cases, requirements, surfaces and defects on the other — and resolve every item on both sides into…
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
Bring back Claude Code sessions, workflows and subagents that a crash interrupted, with the context they already had instead of restarting them from scratch. Use after a terminal, Ghostty or machine crash, a force-quit, or a jetsam/out-of-memory kill took sessions down, and whenever a previous recovery cold-started…
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
report: turn what a Claude session already worked out into a designed, cited report — one self-contained HTML page that reads as a rich scrolling document and paginates to a clean A4 PDF, plus a one-page TLDR. Ships three readings of the same argument over one claim ledger (Primer / Brief / Technical) that the reader…
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
Locate, parse, and resume work from past sessions across any agent CLI (Claude Code, Antigravity/AGY, Cursor, Codex, Grok, or repo workspaces). Discovers sessions by query, ID, path, or recency, extracts the 6D takeover state (original goal, terminal errors, modified files, config keys, decisions, and immediate next…
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
Portfolio-wide orchestrator across every project in /Dev — the layer above ship-fleet:ship-fleet, and the system prompt for a standing master-orchestrator agent/daemon. Reads the portfolio manifest (/Dev/ARMADA.md) and /Dev/CLAUDE.md, verifies freshness against git, then surveys, plans, routes, and dispatches. Route.
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
End-to-end feature delivery conductor: one feature from rough idea to merged, cross-family-verified production code by conducting the shipyard stages - intake, triage, plan in parallel with design, work, the deferred loop, shipyard:gap-fix, acceptance-e2e, verify (the only path to Done), and a fail-closed merge with a.
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
Backlog-wide feature-delivery orchestrator: surveys everything remaining in a repo's pipeline (specs, plans, briefs, mocks, worktrees, deferred follow-ups), writes ORCHESTRATOR.md and a visual hierarchy before any execution, then conducts dependency-ordered ship-feature:ship-feature runners under a global agent budget.
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
The shipyard: the feature-delivery stage skills: intake (rough idea to briefs, with trawl:trawl ideation and Dossier research), triage (readiness with the Sentinel review and decision-gate assumptions), plan (committed plans plus a comprehensive test strategy), design (all-platform interactive mocks gated by…
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
Score 0-10 whether NOW is a good moment to compact a coding-agent session, and say why in one line. Scores the BOUNDARY rather than the token count: an open tool chain or a half-finished edit is a hard zero however full the window is, which is the signal that decided 98.07% of holds across 1,089 measured turns. Reads…
★not rated 2 4d agoA
tokens not measured
originalMIT
Report status as two living HTML pages instead of a long chat message. One project writes .status/project.json; the skill renders that project's STATUS.html, derives its row on the portfolio dashboard at /Dev/STATUS.html, opens the page and replies in two lines. Built from 2,400 real status reports, so the zones are w.
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
Count what a tracker board claims against what the codebase actually holds, card by card, and move each one to the column its evidence supports. Reads every comment and attached image and rebuilds the numbered requirement list from the ticket BEFORE opening any completion record or diff, because a diff read first…
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
Verify and clean up after a finished agent session — a Gemini or other cheaper-model run — without re-doing its work. Reads the session's own transcript against the repository it changed, and lands every claim the session made in exactly one of eight classes with an exit code that blocks a report which lost an item.…
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
Run a complete UI test campaign against an application and leave behind a living evidence page — coverage, requirements, user-flow storyboards, screenshots, component atlas and defects in one browsable surface where every row has a stable id somebody can point at. Reads the project first — Overview, PRD, feature specs.
★not rated 2
changed 5d agoA
tokens not measured
originalMIT
Trawl: evidence-grounded parallel divergent ideation for coding agents — cast a wide net over the idea space, then converge on something you can ship. Successor to the original adhd skill. Isolated framed branches, a frozen textbook baseline as both ban-list and boss gate, a balanced frame portfolio with fit floor and.
★not rated 2
changed 1mo agoA
tokens not measured
originalMIT
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: