The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.run in bin/hunter.py to execute GitHub CLI (gh) commands.
The implementation follows security best practices by passing arguments as a list rather than a shell string, effectively preventing shell injection vulnerabilities.
The commands are scoped to repository searching (gh api) and issue listing (gh issue list).
[INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources (GitHub API) which could theoretically contain malicious instructions.
Ingestion points: Repository metadata and issue titles are ingested via gh api and gh issue list in bin/hunter.py.
Boundary markers: None explicitly used in the output dossier.
Capability inventory: The skill uses subprocess.run to call the gh tool; however, this is not used to execute the ingested data as code.
Sanitization: The use of list-based arguments in subprocess.run ensures that ingested repository names or issue strings cannot trigger command injection.
[EXTERNAL_DOWNLOADS]: The skill references the author's GitHub repository github.com/jackjin1997/ClawForge. No automated downloads or remote script executions from unknown sources are performed at runtime.

oss-hunter