Skills
skills/paultyng/skill-issue/set-session-context/Gen Agent Trust Hub

set-session-context

Pass

Audited by Gen Agent Trust Hub on Jun 20, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses git symbolic-ref and gh pr view to fetch repository and pull request information. These are standard commands for development environments and are used appropriately for context gathering.
  • [PROMPT_INJECTION]: There is a surface for indirect prompt injection as the skill incorporates untrusted data from Git branch names and GitHub PR titles into the /rename command.
  • Ingestion points: Output of git and gh commands processed in SKILL.md.
  • Boundary markers: Not present; external strings are not wrapped in delimiters to separate them from instructions.
  • Capability inventory: Impact is limited to the platform's UI commands /rename and /color.
  • Sanitization: Branch names undergo basic humanization (stripping prefixes and replacing hyphens), but PR titles are used directly as provided by the GitHub CLI.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 20, 2026, 05:34 PM
Security Audit — agent-trust-hub — set-session-context