The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to perform repository management tasks, including listing files, searching with grep, and executing a local Python script scripts/learning-db.py to retrieve learning data.\n- [PROMPT_INJECTION]: The skill contains a surface for indirect prompt injection because it ingests and analyzes the content of other agent and skill files to identify structural gaps and propose improvements.\n
Ingestion points: Repository files located in agents/ and skills/ directories and the output from scripts/learning-db.py.\n
Boundary markers: Absent. The skill lacks instructions to treat ingested file content as untrusted or to ignore instructions embedded within those files.\n
Capability inventory: The skill is granted significant permissions, including the ability to run shell commands via Bash and modify files using Edit and Write.\n
Sanitization: Absent. There are no mechanisms described for sanitizing or validating the data extracted from target files before it is used in the upgrade process.

agent-upgrade