commit-push
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes repository data, creating a potential surface for indirect prompt injection if malicious files are present in the working directory or Git history. 1. Ingestion points: The skill is instructed to inspect Git changes, outgoing commits, and hook outputs in SKILL.md. 2. Boundary markers: No explicit delimiters or instructions to ignore embedded content are provided in the skill files. 3. Capability inventory: The skill invokes Git commands and triggers repository-defined hooks and regeneration scripts as described in SKILL.md. 4. Sanitization: No sanitization or validation of the files being committed or the output of hooks is defined.
- [COMMAND_EXECUTION]: The skill directs the agent to execute shell commands for Git management (commit, push) and to run repository-specific hooks and verification tasks. The instructions include security-positive directives to never bypass hooks using flags like --no-verify.
Audit Metadata