The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local Python scripts (e.g., lint_wiki.py, extract_page_claims.py, resweep_planner.py) using the Bash tool. These scripts are part of the plugin's own distributed codebase ('vendored' scripts).
[EXTERNAL_DOWNLOADS]: The skill performs network operations via the cogni-claims and knowledge-fetch sub-skills to retrieve live web content for claim verification and research. This behavior is documented, purposeful, and involves established vendor services.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection because it reads and processes user-provided wiki content.
Ingestion points: Wiki pages located at wiki_path are read for linting and claim extraction.
Boundary markers: The skill does not explicitly define markers in this orchestrator level, though it uses structured JSON for script communication.
Capability inventory: The skill can execute shell commands via Bash and dispatch other skills.
Sanitization: Parsing and extraction are handled by internal Python scripts.

knowledge-refresh