The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process potentially untrusted external data such as repository issues, product requirement documents (PRDs), and chat contexts in Step 2 ('Gather the minimum credible evidence'). When combined with the skill's capabilities to execute shell commands and modify files via the allowed tools, this allows malicious instructions hidden within those inputs to potentially influence the agent's planning output or behavior.\n
Ingestion points: Untrusted data enters the agent context from source materials like repo/issues, prd/spec, gdd/playtest, and chat-context as specified in SKILL.md.\n
Boundary markers: The instructions do not define explicit delimiters or 'ignore' instructions for the processed content to prevent the agent from obeying embedded commands.\n
Capability inventory: The skill has access to Bash, Read, Write, Edit, Glob, and Grep tools, providing a broad capability surface.\n
Sanitization: No evidence of input validation, escaping, or sanitization was found for the incoming external data.

task-planning