codeck-outline
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes multiple local shell scripts (e.g.,
init-room.sh,status.sh,scan-materials.sh) located within specific plugin or home directory paths (~/.claude/skills/codeck,~/.codex/skills/codeck, etc.) to initialize the project environment and manage tasks. - [INDIRECT_PROMPT_INJECTION]: The skill scans and reads arbitrary local files (text, PDF, DOCX, CSV) to extract key information and generate slide structures. This creates a surface where instructions embedded within those files could potentially influence the agent's narrative planning.
- Ingestion points: Scans the current user directory and reads found files using the
Readtool. - Boundary markers: No specific delimiters or "ignore instructions" warnings are applied when processing these materials.
- Capability inventory: Includes file system writes to the project directory, shell command execution for environment setup, and web search functionality for research.
- Sanitization: No sanitization or validation of the content extracted from project files is documented.
Audit Metadata