The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a local utility script scripts/check-output.py used by the agent to validate the structure of its own generated deliverables. The script is limited to reading local files and checking for specific section headers.
[INDIRECT_PROMPT_INJECTION]: The skill implements a workflow where scripts/check-output.py processes file content, creating a data ingestion surface. 1. Ingestion points: scripts/check-output.py reads content from a file path provided via command-line arguments as described in SKILL.md. 2. Boundary markers: No explicit delimiters or safety instructions are defined for the input file processed by this script. 3. Capability inventory: The script uses path.read_text() in scripts/check-output.py to access local files. 4. Sanitization: No sanitization is performed on the ingested text; however, the script's logic is restricted to non-interpreting string membership checks against a static list of required headers.

sequencing-toolkit