The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill monitors and processes data from local project files to determine its routing logic, creating a vulnerability surface for instructions embedded in file metadata or content. * Ingestion points: The skill reads the status field from journal/NN_<short_name>.md and uses the NN_<short_name> stem from the filesystem to construct test filenames. * Boundary markers: Absent; the instructions rely on the LLM's ability to locate and interpret the 'approved' status within the design note markdown. * Capability inventory: The skill can create new files in the tests/ directory and dispatch control to other specialized subskills. * Sanitization: The skill does not explicitly define sanitization for the stem strings or the contents of the journal files being processed.

test-ml-pipeline