Odin Skills

Use when asked to create or refresh a personal mode skill and open a reviewable PR. Not for shared repo skills or work that skips human approval.

Use when asked to deduplicate a skill tree, fold overlapping skills, or cut the skill count: analyze, gate per family, then fold. Not for prompt-doctrine cascades: use cascade-dedup.

Use when the user names one book, course, paper, or source document and asks to distill it into a reusable skill. Not for a folder of sources: use map-corpus.

Use when prompt-doctrine is duplicated, drifted, or conflicting across output-style embeds and external harness AGENTS files. Not for editing the canonical baseline itself.

Use when a user asks to scan, audit, or validate a skill for security issues. Not for source-code or infrastructure review: use security-review. Not for remote-system changes.

Use when asked to run /model-retuning to retune a skill corpus for a new model, measurement-first. Not for static audits: use deslop.

Use when a user requests an independent Codex or Gemini review of uncommitted code, a branch diff, or a specific commit. Don't use for unconfirmed or unattended review sends, or source mutation.

Use when the user runs /skill-benchmark to score skills, compare models, or gate a skill release. Modes: score (default) and gate. Not for editing skills: use agent-surface-forge.

Use when asked to improve, optimize, rewrite, tune, or port a prompt, skill, or tool description. Also for building prompt evals or auditing prompt text for dated instructions.

Use when asked to generate a personal taste skill from local evidence. Not for applying an existing taste register: use the user-private spine skill.

Use when the user wants to compile taste and "this feels wrong" signals into machine-consumable tokens, rules, forbidden combinations, and examples. Not for remote or irreversible changes.

Use when a publisher wants a discoverable, integrity-protected agent-skill catalog served from a domain they control. Not for remote mutation or deployment. The publisher deploys.

Use when authoring or restructuring an agent-consumed document, a SKILL.md or skill directory, or deciding a skill split-or-monolith disclosure question. Not for prose style: use unslop.

Use when asked to audit or repair agent surfaces (plugins, agents, skills, CLAUDE.md/AGENTS.md, docs, prompts, commands, hooks) or improve one skill at depth. Not for agent grading: use skill-doctor.

Use when skills must move between plugins, or plugins be created, merged, or retired. Not for reordering a listing, use reorder. Not for retiring a code path, use deprecate-and-migrate.

Use when asked to create a local agent-plugin directory tree or marketplace package. Not for remote, credential, publish, deploy, or irreversible changes.

Use when a human invokes the release gate to prove every ODIN plugin shares one canonical version. Not for editing release metadata or remote, credential, publish, deploy, or irreversible changes.

Use when asked to review a plugin for marketplace readiness via a read-only audit of published quality gates. Not for reviewing a PR or code diff: use review.

Use when the user asks to discover, generate, refresh, or validate skills from the Trail of Bits Testing Handbook or appsec.guide. Not for tasks that require source or remote-system changes.

Use when changing the wording of a rule in a skill, prompt template, or agent instruction where the change is meant to alter model output.