git-nuget-release-notes
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses local shell commands, specifically
gitandrg(ripgrep), to identify project files, analyze git history, and inspect repository metadata. These commands are limited to local read operations and standard development workflow tasks. - [INDIRECT_PROMPT_INJECTION]: The skill processes external data from git logs and file diffs that could contain malicious instructions embedded by an attacker.
- Ingestion points: The agent is instructed to read commit subjects, bodies, and file diffs (SKILL.md, Step 5) which are external and potentially untrusted inputs.
- Boundary markers: The skill lacks formal delimiters but uses instructional constraints to separate the data analysis phase from the output generation phase.
- Capability inventory: The skill can execute shell commands (
git,rg) and perform file system writes to update.nuget/*/PackageReleaseNotes.txtfiles. - Sanitization: Instructions explicitly command the agent not to "dump commit subjects verbatim" and to "describe only surviving package outcomes," which functions as a logical filter against direct instruction passthrough.
Audit Metadata