Conciseness

The body is dense and assumes Claude's competence (no explanation of what SAML or STS is), with commands like 'unset AWS_ACCESS_KEY_ID AWS_SECRET_ACCESS_KEY AWS_SESSION_TOKEN' given without padding. Not 5: the 'Use This Skill When' section duplicates the frontmatter description almost verbatim, and 'Do not document or commit live access keys' is guidance Claude does not need — minor trimming is possible. Not 3: there is no unnecessary conceptual explanation, only these small redundancies.

Actionability

Fully executable, copy-paste-ready commands cover the common cases: 'aws sts get-caller-identity --profile <profile-from-login-output>', 'which -a aws', the full 'aws rds describe-orderable-db-instance-options' invocation with --query/--output/sort pipeline, and a parameterized saml2aws example. Not 4: the commands are complete and specific enough that placeholders are the only substitution needed, with no missing key details.

Workflow Clarity

The 'Invalid XML' section gives an ordered diagnostic sequence ('Typical causes (check in order)') covering credential source, env vars, multiple binaries, proxies, and SIGTSTP, and running 'aws sts get-caller-identity --profile ...' functions as the verification checkpoint. Not 5: the validation loop is implicit rather than an explicit 'verify identity, then proceed' step, and the RDS section reacts to 'Cannot find version' failures without an explicit retry loop. Not 3: the checkpoint (get-caller-identity) and a check-order sequence are clearly present, so validation is not merely implicit-and-missing.

Progressive Disclosure

No bundle files exist, and the single-file body is well organized into tight, navigable sections (Goal, Use When, post-login, Invalid XML, Region/IaC, saml2aws pinning, What not to do). Not 5: the body is ~79 lines — above the under-50-lines threshold where a single well-organized file earns full marks — and material like the RDS engine-version recipe could live in a reference file. Not 3: nothing is buried or inlined that clearly belongs elsewhere; structure is clean.

Specificity

The description names multiple concrete actions and artifacts: 'using the AWS CLI against a corporate AWS account after SAML/OIDC federation', 'selecting profiles or regions for Terraform or CI/CD stacks', and 'validating RDS engine versions with describe-orderable-db-instance-options' — comprehensive coverage within its niche. Not 4: unlike 'minor gaps in coverage', the triggers span the login tools, failure symptoms, IaC contexts, and the specific RDS API.

Completeness

Explicit, concrete 'when' triggers are present ('Use when using the AWS CLI... when STS fails... when selecting profiles or regions...'), but no 'what' is ever stated — the description never says what the skill does (e.g., diagnose credential/profile/region misconfiguration). Not 2: unlike the vague anchor 'Use when working with documents', the when-clauses are numerous, specific, and embed action verbs, so it sits above an only-when/what-absent score of 2. Not 4: a 'what' statement is absent, not merely implicit.

Trigger Term Quality

Good keyword coverage including tool names users actually say ('saml2aws', 'aws sso login', 'Terraform', 'CI/CD', 'RDS engine versions'). Not 5: it paraphrases the error users would paste ('STS fails with invalid XML or empty responses' rather than the literal 'Unable to parse response' or 'b\'\'' the body documents), and omits synonyms like 'federated login' or 'wrong account'.

Distinctiveness Conflict Risk

Clear niche with distinct triggers — corporate SAML/OIDC federation with AWS CLI, STS invalid XML errors, and RDS orderable-engine validation — minimal overlap with generic AWS skills. Not 4: the federation/STS-error/RDS-validation combination is specific enough that 'minor overlap risk with closely related skills' understates how targeted these triggers are.