The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes content from various external files to guide its implementation and commit actions, creating a surface for indirect prompt injection.\n- Ingestion points: Reads project-specific files including requirement.md, design.md, tasks.md, coding-rules.md, CLAUDE.md, and AGENTS.md (SKILL.md). It also ingests feedback files (review, test, evaluate) and GitHub issue data via gh issue view (SKILL.md).\n- Boundary markers: The instructions do not specify any delimiters or safety prompts to isolate external content from the agent's core instructions.\n- Capability inventory: The skill can read local files, write/modify code files, and execute shell commands including gh issue view and git commit (SKILL.md).\n- Sanitization: No sanitization, escaping, or validation of external content is described before the data is used to drive code implementation or commit message generation.\n- [COMMAND_EXECUTION]: The skill utilizes command-line tools to interact with the development environment and external services.\n- Evidence: Fetches issue data from GitHub using the gh issue view command (SKILL.md).\n- Evidence: Performs git commit operations based on conventions found in project documentation (SKILL.md).\n- Evidence: Uses file discovery commands like find to locate workflow and coding rules files (SKILL.md).

spec-code