quick-commit
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill contains instructions that explicitly override default agent safety behavior regarding user confirmation. It mandates that the agent "Never ask, never suppose," "do not request confirmation," and "Start the Flow at step 1 immediately. Do not ask for a task."
- [COMMAND_EXECUTION]: The skill uses the shell to execute
git commit. By combining this with the suppression of user confirmation, it allows the agent to perform write actions on the local repository without human-in-the-loop oversight. - [INDIRECT_PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests untrusted data from the local environment and uses it to drive agent behavior (generating the commit message) which is then immediately used in a command.
- Ingestion points:
git diff --cached(SKILL.md, Flow step 3). - Boundary markers: Absent. There are no instructions to the model to ignore or delimit potentially malicious instructions found within the staged code changes.
- Capability inventory:
git commit(SKILL.md, Flow step 5). - Sanitization: Absent. The diff content is processed directly by the LLM without filtering or validation.
Audit Metadata