The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external simulation data (JSONL/CSV) to inform the agent's recommendations for code changes (the 'Patch Plan').
Ingestion points: The script scripts/evaluate_rtp_runs.py reads data from files provided via the --input argument.
Boundary markers: No explicit delimiters or instructions are provided to the agent to treat data within the simulation files as untrusted or to ignore instructions embedded within them.
Capability inventory: The agent is instructed to deliver a 'Patch Plan' containing exact file and function edits and verification commands based on the analysis.
Sanitization: The skill lacks mechanisms to sanitize or validate the content of the simulation data to prevent it from influencing the agent's subsequent logic or recommendations.

rtp-optimizer