ai-create-prd
Warn
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes multiple shell commands to manage the Obsidian vault and its version control state. It uses
git rev-parse,ls, andbasenamefor directory resolution, and a composite commandgit add -A && git commit && git pushto synchronize changes with a remote server. - [DATA_EXFILTRATION]: The workflow involves pushing the contents of the user's Obsidian vault ($HOME/Documents/obsidian/obsidian) to an external Git repository. The instruction to use
git add -Astages all files in the vault, which could lead to the unintended upload of sensitive personal data or secrets stored in unrelated notes within the same vault. - [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it reads content from existing vault files (such as index.md and previous PRDs) to maintain the document graph and then performs file system and shell operations.
- Ingestion points: engineering/index.md, engineering//index.md, engineering//workplans//index.md
- Boundary markers: Absent
- Capability inventory: Bash shell execution (git, ls, basename), file write, and file edit tools
- Sanitization: None detected
Audit Metadata