release

Warn

Audited by Gen Agent Trust Hub on Oct 11, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes several shell commands to manage the release process.
  • Runs python3 scripts/changelog_draft.py to generate changelog entries.
  • Executes git tag and gh release create to perform versioning and publishing operations.
  • Uses ls and conditional shell checks to verify the existence of overlay directories.
  • [DYNAMIC_EXECUTION]: The skill implements a dynamic instruction loading mechanism through the 'Overlay' feature.
  • It checks the $POWERSUIT_OVERLAY environment variable and dynamically reads files from that path (README.md, sdlc/release/).
  • The instructions explicitly state: "Where the overlay and this file disagree, follow the overlay," allowing instructions loaded from a computed environment path to override the core skill logic at runtime.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection by ingesting and obeying external instructions.
  • Ingestion points: Files located in $POWERSUIT_OVERLAY/sdlc/ and $POWERSUIT_OVERLAY/sdlc/release/, as well as commit messages processed by the changelog script.
  • Boundary markers: The instructions include a specific rule to follow the overlay in case of disagreement, which prioritizes the external content over the original skill definitions.
  • Capability inventory: The skill has the capability to execute Python scripts, perform Git operations, and interact with the GitHub CLI (gh).
  • Sanitization: There is no evidence of sanitization or validation of the content loaded from the overlay files before the agent adopts them as instructions.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Oct 11, 2026, 12:50 PM