The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external markdown plan documents through the decision_tree_extractor.py and question_generator.py scripts. These documents are untrusted inputs that could contain malicious instructions intended to manipulate the agent's behavior during the interrogation process.
Ingestion points: Plan markdown files read by scripts/decision_tree_extractor.py and scripts/question_generator.py.
Boundary markers: No explicit delimiters or instructions are used to isolate the content of the plan from the agent's system instructions.
Capability inventory: The skill encourages the agent to use filesystem tools like grep, find, and file reading to verify information.
Sanitization: The scripts do not perform sanitization of the input text before processing.
[COMMAND_EXECUTION]: The skill instructions in SKILL.md and references/forcing_question_patterns.md explicitly direct the agent to execute shell commands (e.g., grep, find) to explore the local codebase as a means of answering questions. This provides a mechanism for the agent to interact with the filesystem based on the content of the documents being analyzed.

grill-me