The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill reads and processes user-controlled project files such as .devflow/index.md, state.md, and brief.md. \n
Ingestion points: The agent is instructed to read various files within the .devflow directory to determine state and context. \n
Boundary markers: The instructions do not specify any delimiters or safety warnings for handling the content read from these files. \n
Capability inventory: The skill can read and write files within the project repository and contains utility functions for spawning subprocesses (scripts/runtime/common.mjs). \n
Sanitization: There is no evidence of sanitization or filtering of the content read from these files before it is used by the agent to make decisions or generate output. \n- [COMMAND_EXECUTION]: The runtime utility scripts/runtime/common.mjs contains a runChild function that uses node:child_process.spawn to execute external programs. \n- [DYNAMIC_EXECUTION]: The skill includes JavaScript files (scripts/run.mjs, scripts/runtime/common.mjs, scripts/runtime/integrity.mjs) that are executed at runtime to perform integrity checks and manage execution flow.

sketch