Skill Claude CodeCodex
Executes one task of an already approved plan, red to green to refactor, as one logical change with a clean checkpoint that says nobody has reviewed it yet. Trigger for "implement task 3", "build this", "make the plan's next step work", "write the code for this spec". Not for deciding what to build — use /ai-spec…
Skill Claude CodeCodex
Grills a written specification from outside it: at most ten questions per round, one at a time, each anchored to the sentence it attacks and decided by executing the commands behind it. Trigger for "challenge this spec", "attack this decision", "is this spec true", "grill this before I approve it". Not for judging a…
Skill Claude CodeCodex
Reads one specification through five lenses that never see each other, then has them read each other anonymously to refute what does not hold and to name what every one of them missed, in one pass whose verdict the author writes into the specification. Trigger for "council this spec", "read this from several angles"…
Skill Claude CodeCodex
Walks one request through the governed cycle by loading each stage's own skill body and following it — research, spec, challenge, council, then a brief a person reads. After that person says go, ai-cycle build runs the second half: build, review, verify, security, audit, ship. Trigger for "run the whole cycle on…
Skill Claude CodeCodex
Finds the root cause of broken behaviour and names it at file:line, then writes the check that fails for that reason before changing anything. Also resolves merge and rebase conflicts by intent rather than by taking a side. Trigger for "it's not working", "this used to work", "I'm getting an error", "CI is failing"…
Skill Claude CodeCodex
One gateway with four routes for creating, extending or redesigning a web, mobile, native or CLI experience: shape the work, build the system, opt into imagery, and verify the rendered result rather than the declared CSS. Trigger for "design this screen", "build the design system", "make this responsive", "check the…
Skill Claude CodeCodex
Answers questions about this repository by reading it, anchored to file:line, and tours an unfamiliar area for somebody who has just arrived. Trigger for "where does X live", "how does this work", "why does it do that", "what depends on Y", "walk me through", "map this module", "trace this import chain", "onboard me".…
Skill Claude CodeCodex
Runs one request through the whole governed cycle in a single pass — research, spec, challenge, council, build, review, verify, security, audit, ship — and stops only to hand a person a tested thing to try. Your invocation is the standing approval; nobody else steps in until the end. The bar is the green gate plus an…
Skill Claude CodeCodex
Saves a finding that took real time to reach — a non-obvious behaviour, an integration trap, a workaround and the reason it is needed — as committed markdown, stamped with the commit and the file patterns it describes so that it can be detected as stale later. Searches those notes too. Trigger for "save this", "note…
Skill Claude CodeCodex
Turns an approved spec into a numbered task list where every task names one file, one check that fails, and how to undo it. When the spec adds anything deployable, the plan must carry a CI/CD task and an observability task; they are not optional. Trigger for "break this down", "what tasks do we need", "let's start…
Skill Claude CodeCodex
Reports a reproducible fault in this framework through ai-eng report issue: a payload closed to nine allow-listed fields, scanned for machine paths, personal data and secrets, written as a local gitignored draft, and shown as the exact bytes with their SHA-256 before anybody confirms anything. Trigger for "report this…
Skill Claude CodeCodex
Finds evidence from outside this repository and reports it with numbered citations, or marks a claim [unsourced] and leaves it marked. Uses only the tools the client actually has: the local floor is always on, web search / Tavily / Exa run only when the client configured them, and NotebookLM deep research runs only…
Skill Claude CodeCodex
Judges a diff the way a staff engineer would: does it do what it claims, is it correct at its boundaries, is it safe, is it tested, and is there a smaller version of it. Never claims the result of a mechanical gate — those ran in CI and it says so. Trigger for "review this", "any issues with this", "is this…
Skill Claude CodeCodex
Reasons about trust boundaries, authentication, data, dependencies, skills and supply chain, runs the scanners this repository already pins, and reports PASS, FAIL or INCOMPLETE with the command behind each one. Trigger for "threat model this", "is this auth safe", "review the trust boundary", "audit these…
Skill Claude CodeCodex
Lands the work: one commit per change, the changelog entry, a pull request that opens with what changed in plain words, the production-ready boxes ticked with the command that ticked each, the work item closed by a keyword in the body, and the branch cleaned up afterwards. Trigger for "ship it", "open a PR", "I'm…
Skill Claude CodeCodex
Writes the governed record of a decision before code exists: evidence, the problem, three real options, one recommendation and self-challenge, assumptions, unresolved risks, observable examples and the authority for proceeding. Trigger for "let's add", "how should we handle", "what's the best approach", "I'm thinking…
Skill Claude CodeCodex
Runs the gate and the security lane and ticks each production-ready box beside the command that ticked it, or walks a spec's examples and marks each one against a real command. Trigger for "verify this", "is it ready", "tick the boxes", "does it do what the spec said", "check the acceptance criteria". Not for judging…
Skill Claude CodeCodex
Turns a plan — one written here, one pasted from a chat, one another agent produced — into the review surface the human actually reads: the plan's own Markdown, carrying visual blocks the renderer turns into diagrams, file maps, decision tables and open questions, handed over as one page whose link opens in the…
Skill Claude CodeCodex
Turns a finished work unit — a branch, a commit range, a PR, the tasks of one plan — into the record a reviewer reads before the raw diff: one page whose file-tree and diff excerpts come from git itself, whose narrative says what changed and why, and whose link opens in the browser beside the digest of the spec it…
Skill Claude CodeCodex
Writes technical documentation for this repository: a README, a wiki page, product documentation, API docs or a technical post. Applies the framework's single writing standard (.agents/skills/ai-report/references/documentation-writer.md, spec 039) and verifies every document against the tree. Trigger for "write the…
Instructions file CodexOpenCode
Instructions for arcasilesgroup/ai-engineering, covering how work happens here, the twelve rules, what this project is, the shape of the tree and the two contracts that must not bend.
Instructions file
Instructions for arcasilesgroup/ai-engineering, a project described as: Turn any repo into a governed AI workspace. Quality gates, security scanning, and risk management — enforced locally via git hooks. Works with Claude Code, GitHub Copilot, Cursor, Gemini & Codex.