The Agent Skills Directory

COMMAND_EXECUTION (LOW): The skill is designed to interact with the system via the codex CLI tool. It uses Python wrappers to execute shell commands like codex exec. While this is the primary purpose of the skill, it represents a capability that could be abused if the agent is influenced by malicious instructions.
PROMPT_INJECTION (LOW): Identified as a Category 8 (Indirect Prompt Injection) vulnerability. The skill ingests untrusted data from project files (specifically CLAUDE.md) to generate a new documentation file (AGENTS.md) intended for use by another AI agent.
Ingestion points: The claude_parser.py and project_analyzer.py modules read local project files and directory structures.
Boundary markers: The provided code does not show explicit delimiters or 'ignore' instructions when interpolating parsed content into the generated AGENTS.md templates.
Capability inventory: The skill possesses the ability to execute shell commands via codex_executor.py and write files to the project root.
Sanitization: No evidence of input sanitization or validation was found in the parsing logic, allowing potential malicious instructions in a project's CLAUDE.md to be reflected into the AI-facing AGENTS.md file.

codex-cli-bridge