The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests codebase files and external research to generate implementation plans, creating a vulnerability surface for indirect prompt injection.
Ingestion points: SKILL.md Step 1 ("Read all mentioned files immediately and FULLY") and Step 3 (spawning research tasks via codebase-locator and codebase-analyzer agents).
Boundary markers: The instructions do not specify delimiters or safety warnings to isolate ingested codebase content from the agent's plan generation logic.
Capability inventory: The skill writes plan files to the thoughts/ directory and generates shell commands for Git operations and testing frameworks.
Sanitization: There is no evidence of sanitization or validation of the content read from the codebase before it is interpolated into the generated plans.
[COMMAND_EXECUTION]: The skill generates and suggests shell commands for local Git and testing operations.
Evidence: SKILL.md and template.md provide templates for git commit, git checkout ., git stash, and test runner commands like npm test and pytest.

tdd-planning