Content
63%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 content is a well-sequenced, largely executable recon workflow with a strong worked scenario and output template. Its main weaknesses are padded sections re-explaining concepts and tools Claude already knows, and a progressive-disclosure failure: the bundled `scripts/agent.py` and `references/api-reference.md` are never referenced from the body, leaving them orphaned.
Suggestions
Add a short 'Automation & References' section pointing to `scripts/agent.py` for automated collection (with its CLI usage) and `references/api-reference.md` for function-level and API details, so the bundle files are discoverable.
Remove or drastically trim the 'Key Concepts' table and 'Tools & Systems' section — Claude already knows what OSINT, certificate transparency, Shodan, and Amass are; keep only target-specific or non-obvious usage notes.
Insert explicit validation checkpoints between steps (e.g., 'resolve and dedupe all subdomains before Shodan/IP correlation; confirm API keys respond before relying on paid sources') to close the workflow's validation gaps.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The workflow steps themselves are lean and command-driven, but the "Key Concepts" table (defining OSINT, passive/active reconnaissance, Certificate Transparency, Google Dorking, Shadow IT) and the "Tools & Systems" section explaining what Amass, Shodan, and theHarvester are re-teach concepts and tools Claude already knows. This matches the level-3 anchor 'Mostly efficient but includes some unnecessary explanation or could be tightened'; it is not severely padded (level 1-2) since the bulk of the body is dense, useful guidance. | 3 / 5 |
Actionability | Most guidance is copy-paste ready: `amass enum -passive -d target.com -o amass_subs.txt`, the crt.sh curl|jq pipeline, `dig axfr @ns1.target.com target.com`, `shodan search "ssl.cert.subject.cn:target.com"`, and specific GitHub dorks. A few steps remain high-level hints ("Use `hunter.io` to determine the email format", "Review employee social media profiles"), matching level 4 'Mostly executable guidance; concrete code or commands with minor gaps' rather than the fully-executable coverage of level 5. | 4 / 5 |
Workflow Clarity | The five-step workflow is clearly sequenced (DNS → infrastructure → personnel → leaks → tech stack) with a consolidation step and a concrete output-format template that serves as the end-state checklist, and the pitfalls warn against scope-creeping into active scanning. Validation checkpoints are implicit rather than explicit (e.g., no 'verify subdomains resolve before Shodan correlation'), which fits level 4 'Clear sequence with most checkpoints present; minor validation gaps' and falls short of level 5's explicit validate-fix-retry loops. This is passive recon, so the destructive/batch cap does not apply. | 4 / 5 |
Progressive Disclosure | The body is well-sectioned (When to Use, Prerequisites, Workflow, Scenarios, Output Format), but it never references the bundle files that actually exist — `scripts/agent.py` and `references/api-reference.md` are orphaned and undiscoverable from SKILL.md, while tool/API detail that belongs in those files is inlined instead. This matches level 3 ('references present but not clearly signaled; content that should be separate is inline') — the internal structure prevents a score of 2, but the complete absence of navigation to real bundle files blocks level 4. | 3 / 5 |
Total | 14 / 20 Passed |