Conciseness

The rules, syntax table, and usage examples are lean, but sections like 'SigV4 signing' (credential resolution order, service/region inference from hostname) and 'How It Works' (re.sub callable mechanics, subprocess internals) explain implementation details the agent never acts on — these belong in or alongside the bundled script. Not 4 because two full paragraphs of signing internals and resolution-order mechanics are clearly trimmable; not 2 because there is no padding explaining concepts Claude already knows.

Actionability

Concrete, copy-paste-ready bash examples covering psql, curl, mysql, docker, and config templating, plus a precise syntax table and actionable troubleshooting (specific stderr formats, 'aws sts get-caller-identity', ASM_EXEC_MCP_TIMEOUT). Not 5 because the skill never states where the asm-exec script lives or how to make it invokable — the references/asm-exec bundle file is not mentioned by path, so the agent cannot fully execute the guidance unaided.

Workflow Clarity

The usage pattern is unambiguous (MUST NOT fetch directly; MUST use {{resolve:...}} via asm-exec), and the troubleshooting section provides a genuine feedback loop (read the stderr cause line first, check credentials/region/permissions, interpret 401 vs. missing secret, raise timeout). Not 5 because there is no pre-flight checkpoint (verify SMA is running or credentials are resolvable before first use) and no guidance on obtaining/invoking the script itself; well above 3 since error recovery is explicit and the single action is clear.

Progressive Disclosure

Sections are well-organized, but the bundle contains references/asm-exec and the body never signals that path — 'How It Works' and 'SigV4 signing' implementation detail is inlined in SKILL.md where a one-level-deep reference to the script (or a short README for it) belongs. Fits anchor 3 ('references present but not clearly signaled; content that should be separate is inline'); not 4 because the sole bundle file is unreferenced and navigable discovery of it depends on guesswork.

Specificity

Names the domain ('AWS Secrets Manager, secret management, credentials, API keys, tokens, and passwords') and two concrete actions ('Prevents AI agents from directly fetching secret values and teaches runtime dynamic references with asm-exec'), but coverage is not comprehensive — resolution backends, config templating, and enforcement hook capabilities are absent. Not 4 because only two actions are listed with minor gaps beyond domain naming; not 2 because the actions are genuinely concrete, not generic.

Completeness

The 'what' is clear (prevents direct secret fetching, teaches runtime dynamic references with asm-exec so plaintext never enters the LLM context window), but there is no 'Use when...' clause or equivalent explicit trigger guidance — the rubric guideline caps this at 3. Not 4 because the 'when' is entirely absent rather than merely imprecise.

Trigger Term Quality

Good natural keyword coverage: 'credentials', 'API keys', 'tokens', 'passwords', 'secret management', 'AWS Secrets Manager' — phrases users would naturally say. Not 5 because common variations like 'secrets', 'environment variables', or tool-specific phrases (e.g. 'get-secret-value') are missing; clearly above 3 since synonyms beyond the bare domain name are present.

Distinctiveness Conflict Risk

A clear niche — preventing agents from fetching AWS Secrets Manager values and resolving them at runtime via asm-exec — with distinct triggers unlikely to fire for unrelated skills. Not 5 because generic terms like 'credentials', 'API keys', and 'passwords' could overlap with other secret-handling or security skills; well above 3 given the highly specific behavioral framing.