Skill Claude CodeCodex
Use before acting on human instructions: separate what they said from what they meant.
not rated 2
changed 2d ago A 20 tokens
Skill Claude CodeCodex
Use when deciding where knowledge goes or reading/writing durable docs: AGENTS.md, .context/, KB, docs/, and work directories.
not rated 2
changed 2d ago A 32 tokens
llm-writing
27Skill Claude CodeCodex
Load before writing or revising human-facing text. Choose words deliberately, ground the piece in the reader's context, and remove default LLM phrasing before the final draft.
not rated 2
changed yesterday A 39 tokens
Skill Claude CodeCodex
Use when validating markdown links or Mermaid diagrams.
not rated 2 yesterday A 12 tokens
qi-layer
29Skill Claude CodeCodex
Use when writing or maintaining AGENTS.md, .context/CONTEXT.md, or CLAUDE.md mirrors: keep intent docs minimal and load-bearing.
not rated 2
changed 2d ago A 34 tokens
Skill Claude CodeCodex
Load on agents that make durable changes to a project's files, so the colocated AGENTS.md and .context/ docs stay true for the next cold agent.
not rated 2
changed yesterday A 36 tokens
reflect
31Skill Claude Code
Use when the user invokes it before you report: checks you delivered what they needed, not only what they literally asked, and routes knowledge capture.
not rated 2
changed 2d ago A 31 tokens
Skill Claude CodeCodex
Use when recovering decisions or context from conversation history: session log navigation, transcript delegation.
not rated 2
changed 2d ago A 21 tokens
shared-dao
33Skill Claude CodeCodex
Use when establishing shared vocabulary: resolve terminology early, one name per concept, vocab files as record.
not rated 2 yesterday A 24 tokens
Skill Claude CodeCodex
Use when a workspace has changes from other agents: safe orientation, preserving unfamiliar files.
not rated 2 yesterday A 21 tokens
Skill Claude CodeCodex
Load when creating structured HTML artifacts to explain and present information.
not rated 2
changed 8d ago A 16 tokens
Skill Claude CodeCodex
Load when work must survive compaction or crashes or be shared with other agents: findings, decisions, and artifacts go in the work directory, not in context.
not rated 2
changed yesterday A 36 tokens
zoom-out
37Skill Claude Code
Use when orienting in an unfamiliar area: fans out parallel explorers across structure, knowledge, and history, then synthesizes one orientation map.
not rated 2
changed yesterday A 31 tokens
Skill Claude CodeCodex
Use when a spawn fails due to sandbox or permission restrictions.
not rated 2
changed 2d ago A 21 tokens
correct
39Skill Claude Code
Use when the human wants the mistakes agents keep repeating fixed for good. Works the KB corrections register: fixes each recurring class at the strongest level that holds (remove the wrong way, an automated check, a tool permission, a written rule last), proves each fix on a real past mistake, and updates the…
not rated 2
changed yesterday A 65 tokens
cold-reader
40Agent Claude Code
Spawn to check text that outlives the work (KB pages, AGENTS.md, docs, PR bodies, prompts) for what a reader who wasn't there can't follow. Pass only the paths and who reads each one: no summary of the work, no reasons for the changes, and no labels from it, even as examples. Returns each place the reader stalls…
not rated 2 yesterday A 89 tokens
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: