ai-modules

Close one completed or parked task. Use when the user asks to finish, mark done, defer, park, drop, or archive a task. Set finished or deferred, move the file to archive, update links, and relint.

Manage the project task backlog as plain markdown files in tasks. Use for broad backlog work including create, list, query, update, triage, implement, audit, finish, defer, archive, lint, split, or repair tasks.

Implement one existing task file end to end. Use when the user asks to build, do, or implement the work described by a task. Edit code and tests, run verification, stamp implemented, and leave audit and finish to sibling skills.

Select and rank eligible live tasks from the project backlog. Use when the user asks what task to work on next, asks Codex to pick or prioritize backlog work, rank open tasks, choose from tasks/, or recommend the next task/action without editing task files.

Create one structured git commit covering the full current working-tree state. Use when the user explicitly asks to commit in the current turn — "commit", "commit this", "commit changes", "make a commit", "write a commit message", "git commit", or an equally direct ask to put the current repo state into git history. Invoke only on that explicit request; a turn that merely finishes other work, saves files, or wraps up without a commit ask leaves committing to the user.

Autonomously drive one task from open or checked to ready through task_check, verifier-approved body repairs, and final mechanical task-lint cleanup. Use when a user asks to auto-fix readiness issues, make a task ready, refresh a stale task against the current codebase, or run an autonomous readiness loop without implementing the task. A task whose premise the code invalidates stops the loop and surfaces deferral as the user's option.

Check one task before implementation. Use when the user asks if a task is ready to build or still valid. Assess structure, scope, focus, complexity, contradictions, and ambiguity, verify the task's premise and described current state against the codebase, surface a stale invalidated task with deferral as the user's option, then stamp ready or checked and report readiness issues.

Create or update a day-grouped, add-only CHANGELOG.md from git commit history. Use when asked to generate, refresh, or maintain a repository changelog. Produces newest-first day sections while keeping existing entries immutable and processes one day at a time to stay within context limits.

Refresh a local git repo back to a clean current default branch: fetch/prune, detect the remote HEAD, fast-forward main/master by --ff-only, delete merged local branches, and offer gated cleanup for upstream-gone, squash-merged, stale, or force-delete branch pruning. Use when the user asks to clean up branches, refresh main, refresh master, delete merged branches, get back to a clean main, prune stale local branches, or tidy local branches after merged PRs. Invoke only on that explicit request in the current turn; noticing stale or merged branches during other work leaves the cleanup to the user.

Cross-agent-harness and cross-OS portability review for runtime artefacts bundled inside skills and plugins. Use when creating, editing, reviewing, or troubleshooting shell scripts, Bash scripts, Python helpers, Node helpers, hooks, Codex hook layering, Claude hook configuration, shared plugin hook scripts, MCP servers, command wrappers, setup or install flows, plugin wiring, agent and subagent definitions, agent frontmatter fields, tool allowlists, read-only agent enforcement, agent TOML generation, skill wording, execution instructions, initialization instructions, configuration instructions, path handling, environment variables, permissions, or provider-specific OpenAI Codex, Anthropic Claude, Cursor, and Gemini CLI behavior that must work across agent harnesses, macOS, and Linux.

Hub and source of truth for the repo-root guardrail documents (CHARTER.md, ARCHITECTURE.md, TESTING.md, SECURITY.md) that keep AI agents anchored to human intent while a project evolves within its declared boundaries. Explains the doc set, the authority hierarchy, the shared format, and how skills consult the docs, suggests which guardrails fit a repository, and drafts one only on explicit user request. Use when the user asks what guardrail docs are, which ones a repo should adopt, to suggest, bootstrap, set up, or draft a charter, architecture, testing, or security doc, how guardrail docs rank against each other or against CLAUDE.md, AGENTS.md, or GEMINI.md, or how agents consult and enforce these docs. For auditing existing guardrail docs against each other and the codebase, route to guardrail_audit.

Create exactly one new task file for one upcoming work item. Use when the user asks to make, add, file, write, or capture a single task or todo. Use task for listing, querying, updating, archiving, linting, or multi task backlog work.

Repair the whole tasks backlog tree in one pass, with an optional autonomous escalation. Use when the user asks to health check, clean up, audit, lint, or autonomously resolve backlog judgement calls. Run the archive inclusive linter, fix mechanical frontmatter, status, location, link, datetime, and provenance issues inline by default, and escalate to auto_shaper_task only on explicit opt-in or confirmed scale.

Audit one implemented or finished task against the actual codebase. Use after implementation or for drift checks before close out. Inspect code and tests, run verification, stamp audited only on clean implemented work, and report gaps otherwise.

Explain one project task in compact what/why/how prose without editing it. Use when the user asks to explain, summarize, walk through, or orient on a task file by path, partial name, or conversation context. If the task reference is missing or ambiguous, activate and ask which task to explain. Not for readiness, next-work selection, implementation, audit, finish, or repair.

Write AI-consumed content using positive, action-oriented language as the primary carrier of every instruction. Use when creating or editing any artifact an AI reads at inference time — SKILL.md files, .mdc rule files, CLAUDE.md/AGENTS.md/GEMINI.md configuration, prompt templates, system prompts, commands, agent and sub-agent definitions, instruction sets, and persona definitions.

Activate this skill whenever the user mentions their wiki, knowledge base, or research notes in any way — including queries that compare, contrast, reference, analyze, or discuss wiki content rather than ask to edit it. Build and maintain a persistent, compounding knowledge base of interlinked plain markdown files. Use when the user asks to create, build, start, or initialize a wiki or knowledge base; add, create, or write wiki pages; query, compare, contrast, reference, or analyze an existing wiki to answer a research or domain question; archive or reorganize wiki pages; or whenever the user names the wiki, the knowledge base, or their notes in the current request even as a passing reference.

Organize AI-consumed content (prompts, rules, skills, commands, agents, system instructions) into pseudo-XML by wrapping each semantic concern in a dedicated tag for role, policy, inputs, and output contract.

Audit and autonomously fix every issue in the wiki — structure, content, splits, links, tags, scaffold drift, and cross-page contradictions — by invoking the `auto_shaper_wiki` agent. Use when the user asks to fix, repair, lint, audit, health-check, clean up, or auto-repair their wiki.

Import a single external resource (URL, file, paper, PDF, transcript, meeting note, internal note, paste) into the wiki with a user-reviewable triage step before any wiki page is written. Use when the user points at such a resource and asks to ingest, import, integrate, digest, absorb, review-before-adding, or propose-then-add it into their wiki; when they want a triage step on a single source rather than a straight ingest; or whenever a named source should be brought in with a propose-then-act front end.

Wrap up the current chat session by mining it for durable knowledge, comparing the findings against the existing wiki, and surfacing both candidate additions and contradictions with concrete reconciliation suggestions. Use when the user asks to wrap up, close, or end a session; to capture, harvest, or persist what we discussed into the wiki; to ingest the session into the wiki; to reconcile the conversation with existing notes; or whenever a research or exploration chat is concluding and produced reusable knowledge worth keeping.

Apply markdown linting compliance and best practices when creating or editing markdown files (.md, .mdc). Covers blank-line rules around block elements, consistent bullet style, fenced code blocks with language identifiers, table alignment syntax, header level progression, list indentation and spacing, and link/image conventions.

Apply formatting standards, code quality rules, structure conventions, linting prevention, and best practices when generating or editing Python code (.py). Covers indentation, quoting, imports, naming, line length, type hints, error handling, logging, data structures, testing, and security practices aligned with common linters (flake8, ruff, pylint).

Apply clippy-aligned Rust practices when writing or editing Rust code (.rs). Covers procedural flow, clippy-driven clarity improvements, minimal imports, Result/Option idioms, fallible builders with project error types, string prefix/suffix handling, borrowing clarity, iteration style, string building, and function signature grouping.

Create executive summaries from documents, reports, and written content. Use when user asks to summarize a document, create an executive summary, condense content, extract key takeaways, or synthesize written materials. Produces structured prose preserving logic and reasoning at 10-15% of original length.

Convert any input text into a Sparse Priming Representation (SPR) — a compact, markdown-structured set of non-overlapping, informationally dense priming statements that allow another LLM (not previously exposed to the source) to reconstruct the original material as completely as possible. Use when asked to create an SPR, produce priming statements, generate a sparse priming representation, or compress content for LLM-to-LLM knowledge transfer.