skills/garrytan/gstack/context-save/Gen Agent Trust Hub

context-save

Pass

Audited by Gen Agent Trust Hub on Nov 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes various system commands (git, ps, date, mkdir) and vendor-specific binaries (e.g., gstack-skill-start, gstack-paths, gstack-slug) to gather project metadata and manage state files.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources that could contain malicious instructions.
  • Ingestion points: Reads git status, git log, and git diff output which can include content from commit messages or file changes controlled by external contributors.
  • Boundary markers: No explicit delimiters are used for the raw git output during the ingestion phase; the content is passed directly to the LLM for summarization.
  • Capability inventory: The skill has access to Bash (shell execution), Write (file system modification), and AskUserQuestion (user interaction).
  • Sanitization: The skill implements a robust bash-side sanitizer for user-provided titles to prevent shell metacharacter injection and path traversal when generating filenames.
  • [DYNAMIC_EXECUTION]: The skill is configured to follow one-time onboarding and consent directives if they are present in the output of the gstack-skill-start tool. This mechanism is restricted to instructions appearing in the direct tool result of that specific command, linked by a SESSION_ID verification step.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 16, 2026, 08:17 PM