Content
75%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is a dense, highly actionable runbook: real discovery, parsing, validation, and pivoting commands organized in a clear numbered workflow with an embedded credential-validation checkpoint. Weaknesses are minor — unexplained internal tooling calls (tfstate_audit, kg_add_node), undefined shell variables in the RDS pivot, and reference-worthy sections (remediation, exemplars) inlined in the main file.
Suggestions
Replace or annotate the environment-specific calls (tfstate_audit, kg_add_node/kg_add_edge) with a note on what tool provides them or standard equivalents, since they are not executable as written.
Define how $USER/$PASSWORD/$AKID/$SK are populated in sections 4-5 (e.g. sourcing from /tmp/tf-secrets.txt) to make the pivot commands fully copy-paste ready.
Move "Defender remediation" and "Known exemplars" into a references/ file and link them from the body to keep SKILL.md a leaner overview.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dominated by lean, copy-paste command blocks and terse lists with almost no padding. It is not 5 because the opening explanation of what a tfstate file contains, the misconfiguration list, and the "Known exemplars"/"Defender remediation" sections carry background Claude largely knows or could infer, adding tokens beyond what is strictly needed. | 4 / 5 |
Actionability | Nearly every section gives concrete executable commands (curl path loop, feroxbuster, aws s3 ls, jq extraction filters, sts get-caller-identity, mysql/psql connect, terraform backend config, BFG). It is not 5 because "tfstate_audit(...)" and "kg_add_node(...)" invoke unexplained external tooling, and section 5 uses undefined $USER/$PASSWORD placeholders; it is well above 3 because the commands are real and complete rather than pseudocode. | 4 / 5 |
Workflow Clarity | A clear numbered sequence 1-6 (discover → parse → topology → validate creds → pivot → promote) with an explicit validation checkpoint in step 4 (sts get-caller-identity "confirmed live cred" if the ARN returns). It is not 5 because there are no explicit validate-and-retry loops or checks that a downloaded state file is valid before parsing, and not 3 because the sequence and one key credential-validation checkpoint are present. | 4 / 5 |
Progressive Disclosure | This is a single-file skill (no references/, scripts/, or assets/ exist) with clear, well-ordered section headers and a one-level-deep cross-skill pointer (aws-iam-enum/SKILL.md). It is not 5 because at ~175 lines, content such as "Defender remediation" and "Known exemplars" could be split into separate reference files to keep SKILL.md a leaner overview. | 4 / 5 |
Total | 16 / 20 Passed |