The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute GitHub CLI commands (gh search, gh api) and a local Python validation script (validate_survey_artifacts.py) as part of its core research and verification workflow.
[COMMAND_EXECUTION]: The instructions include a pattern to generate and execute a small Python one-liner via the shell to calculate a date for filtering search results.
[PROMPT_INJECTION]: The skill processes untrusted data from external sources, which presents an indirect prompt injection surface.
Ingestion points: Untrusted content enters the agent's context through GitHub repository descriptions and summaries retrieved via WebFetch (e.g., in references/evidence-recovery-ladder.md).
Boundary markers: The skill mandates the use of structured markdown artifacts (triage.md, context.md, solutions.md, platform-map.md) to organize and separate research data.
Capability inventory: The skill possesses capabilities to write files, execute shell commands, and perform network requests.
Sanitization: The instructions define a "relevance gate" in Step 4.5 and 4.6 that filters external data based on specific metadata like licenses, repository age, and community traction (stars).
[DATA_EXFILTRATION]: The skill performs network operations via WebFetch to various domains discovered during the research process to retrieve documentation and vendor information.

survey