Conciseness

The body is dense and assumes competence — a feature map, config discovery rules, and a numbered procedure with no basic-concept padding — matching 'Efficient; minor instances of over-explanation that could be trimmed'. It is not a 5 because there is real duplication: "Reference routing" and "How-To Guide" each route the same guide.md and workflows.md content, and the four built-in workflow bullets detail material that the routing section already delegates to @reference/workflows.md.

Actionability

Mostly concrete and executable: exact paths (`.mimocode/mimocode.json`, `~/.config/mimocode/mimocode.jsonc`), commands (`mimo llm-server issue`, `/goal`, `/compose-next`), a copy-paste JSONC snippet, and a 7-step edit procedure. It falls short of 'fully executable' because key specifics are deferred to references — 'Validate the parsed configuration with the narrowest relevant command' never names that command, and the redaction rule lists key patterns but no concrete redaction example.

Workflow Clarity

The configure workflow is a clear 7-step sequence with scope determination, secret redaction, and a final validation-and-report step — 'Clear sequence with most checkpoints present'. It is not a 5 because the validation step has no feedback loop: nothing says what to do when the validation command fails (fix and re-validate), and for an edit-user-config workflow that error-recovery checkpoint is the missing piece.

Progressive Disclosure

The body's structure and signaling are excellent — a topic-to-file routing section with one-level-deep, clearly named references — but none of the routed files (@reference/providers.md, config.md, guide.md, commands.md, permissions.md, mcp-sampling.md, capability-api.md, workflows.md) exist in the bundle; no references/, scripts/, or assets/ directory is present. As written the navigation is broken, which drops it to 'some structure but could be better organized' territory rather than rewarding routing links that resolve to nothing.

Specificity

The description lists multiple concrete actions — "inspect existing config safely, make minimal changes, and verify them", explain features, configure custom API endpoints — with comprehensive coverage of the tool's surface (TUI/CLI commands, keybindings, agent modes, providers, models, authentication, file locations). It matches the anchor 'Lists multiple specific concrete actions; comprehensive coverage' and is not the level below, which requires 'minor gaps in coverage'.

Completeness

It explicitly answers both questions: when via "Use whenever the user asks about MiMoCode itself… Especially trigger when… Also trigger when…", and what via "inspect existing config safely, make minimal changes, and verify them without guessing schema fields or model capabilities". This is the anchor-5 pattern (clear what + concrete trigger phrases); score 4 would require the 'when' to be less explicit, which it is not.

Trigger Term Quality

Trigger terms are comprehensive and include synonyms and exact file spellings users would type: "base URL/baseURL, API key/apiKey, model name or ID, provider, Anthropic Messages API, global/project mimocode.json/jsonc", plus natural phrases like "rendering glitches", "TUI lag", and "how to enter or leave plan mode". This matches the anchor 'Comprehensive coverage of natural terms including synonyms and file extensions' rather than the level below, which tolerates 'a few natural terms missing'.

Distinctiveness Conflict Risk

The niche is clear and product-scoped — every trigger is anchored to "MiMoCode itself", with product-specific tokens (mimocode.json/jsonc, `mimo llm-server`, /compose-next, build/plan/compose modes) that no generic config or provider skill would claim. Distinct triggers with minimal conflict risk matches the top anchor rather than the 'minor overlap risk with closely related skills' anchor.