The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/skillify_check.py executes several system-level commands via subprocess.run, including git rev-parse for repository metadata, bash -n and node --check for syntax validation, and npx for skill registry checks.
[DYNAMIC_EXECUTION]: The scripts/skillify_check.py script includes an optional --run-tests flag that uses subprocess.run to invoke pytest on code within the tests/ directory of a target skill.
[EXTERNAL_DOWNLOADS]: The scripts/skillify_check.py script performs network requests by executing npx skills add --list against repository targets.
[INDIRECT_PROMPT_INJECTION]: The skill distills potentially untrusted session data into code and tests.
Ingestion points: Chat history and session logs via the look-back lens.
Boundary markers: No specific delimiters are implemented in the automated script to isolate session data.
Capability inventory: Includes file writing, syntax validation, and test execution.
Sanitization: The tool does not explicitly sanitize the ingested session data before distillation.

skillify