Conciseness

The body is dense imperative instruction with no concept explanations and no padding — lines like "proposal records goals and scope; spec defines behavior and acceptance; design_doc points to the single formal technical design" compress an entire artifact model into one line. Every sentence is operational and nothing restates knowledge Claude already has, matching the 'lean and efficient; every token earns its place' anchor.

Actionability

Concrete, copy-ready commands are given for the main paths (`comet classic workspace resolve <change-name> --json`, `comet state select`, `comet resume-probe . --stdin --json`, `comet state transition <name> preset-escalate`) plus a route→Skill dispatch table. It falls short of 5 because key executable details are deferred to reference files absent from this bundle — e.g. "fill a CometIntentFrame" and "run `comet task`" have no inline field list, schema, or output example.

Workflow Clarity

A clear numbered sequence (resolve target/workspace → load phase → complete authorized work) with explicit checkpoints: phase Guards, "phase exit checks", decision points that "block dependent work until the user answers explicitly", preset escalation on changed-file thresholds, and recovery procedures. Not 5 because the control flow for any given situation is scattered across four sections and the compressed phrasing requires re-reading to reconstruct the order.

Progressive Disclosure

The structure is exemplary for a complex skill: an overview that delegates ~11 one-level-deep reference files, each tied to an explicit read trigger ("read Ambient Resume in context-recovery.md", "read the chosen preset's escalation rules"). Docked one level because the referenced files do not exist in this bundle — the `comet-classic/reference/...` paths resolve nowhere and §2 inconsistently uses a different style (`[...](reference/scripts.md#task-context-and-artifact-language)`) — so navigation cannot be verified and the skill is unexecutable as shipped.

Specificity

The only capability statement is "Comet Classic workflow entry", which names the domain but gives no concrete actions — the workspace binding, intent routing, phase dispatch, and state recovery the body actually performs are all absent, matching "Names the domain but actions are minimal or generic" (cf. "Processes PDF files"). Not 3 because no concrete action is listed; not 1 because the domain is named and the trigger conditions are operationally precise.

Completeness

The "when" is explicit and excellent — a "Use when..." clause with three concrete trigger conditions — but the "what" is thin: "workflow entry" does not say what the skill does (route intents, bind workspaces, dispatch phases). This is the inverse of the score-3 anchor (clear what, missing when) and below the score-5 anchor (both explicit), landing at "both present, one could be more explicit".

Trigger Term Quality

Good natural coverage for this niche: "invokes /comet-classic", "asks to start or resume Classic", "resume-probe returns auto_resume" are exactly the phrases a user of this system would say. Falls short of 5 because common synonyms like "continue", "resume the active change", or "Classic workflow" are missing.

Distinctiveness Conflict Risk

"Comet Classic", "/comet-classic", and "resume-probe returns auto_resume for one unambiguously recoverable active Classic change" are terms unique to this skill and its sibling comet skills, and the description explicitly scopes itself to the entry invocation. Clear niche with distinct triggers; minimal conflict risk.