The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes several local Python scripts (cap_probe.py, data_api.py, narrative_cap_probe.py) to interface with the Abel API. These scripts are thin wrappers intended to facilitate data retrieval and analysis as part of the skill's primary function.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources, including the Abel API and web search results, which constitutes an indirect prompt injection surface. * Ingestion points: API responses from cap.abel.ai and search results from the web grounding workflow. * Boundary markers: The skill implements a trust hierarchy for sources (Tier A, B, and C) but does not define explicit syntactic delimiters for all ingested data. * Capability inventory: Includes execution of local scripts and web search capabilities. * Sanitization: Employs render_guard.py to filter raw graph identifiers and internal numeric model outputs from the final agent response.
[DYNAMIC_EXECUTION]: Provided Python scripts dynamically adjust the library search path (sys.path) to include a sibling directory (abel-common). This is used to load shared vendor logic and is a standard practice within the vendor's project structure.

abel-ask