The Agent Skills Directory

[SAFE]: The skill implements a robust and transparent workflow for institutional knowledge management. The logic for evaluating, classifying, and integrating PR findings into a project's documentation is well-defined and focused on long-term maintainability.- [COMMAND_EXECUTION]: The skill references a local Python script scripts/palace_manager.py for search and management tasks. While the script itself is not included in the skill package, its usage as described in the CLI reference and search modules is limited to local palace management operations.- [PROMPT_INJECTION]: The skill ingests data from external PR findings and discussions, which creates a surface for indirect prompt injection.
Ingestion points: Finding evaluation logic in modules/capture-workflow.md reads content from PR findings.
Boundary markers: Not explicitly defined in the logic snippets.
Capability inventory: Knowledge capture and semantic search operations via scripts/palace_manager.py.
Sanitization: Absent in the provided logic snippets; the skill relies on structured templates for knowledge storage which provides some inherent structural isolation.

review-chamber