The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes several shell commands to manage the environment, including git status, ls, git branch, and git log. These are used to inventory session activity and verify the state of the repository.
[COMMAND_EXECUTION]: It invokes local tools and scripts such as ~/.local/bin/claude-md-autogen-gate, organvm context sync, and organvm refresh. These tools are part of the vendor's ecosystem and are used to verify and update project metadata.
[INDIRECT_PROMPT_INJECTION]: The skill analyzes files authored or modified during the session to determine their closure status, which introduces an attack surface for indirect prompt injection.
Ingestion points: Reads markdown plans from ~/.Codex/plans/*.md and structured data from data/prompt-registry/prompt-atoms.json to identify status references such as DONE-NNN or IRF-XXX-NNN.
Boundary markers: There are no explicit boundary markers or instructions to ignore potential commands embedded within the session plans being analyzed.
Capability inventory: The skill has capabilities to create and modify files, move plan files to an abandoned directory, and execute shell commands (git, ls, mv, organvm).
Sanitization: Content from the analyzed files is not sanitized or validated before being processed to determine the session's completion state.

closeout