session-wrap-up-agent-skill

Pass

Audited by Gen Agent Trust Hub on Mar 7, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection through its ingestion of untrusted repository data.\n
  • Ingestion points: Data enters the context from git logs and project-specific instruction files like AGENTS.md or CLAUDE.md in SKILL.md (Steps 1 and 2).\n
  • Boundary markers: No delimiters or instructions to ignore embedded commands are used when the agent reads external file content.\n
  • Capability inventory: The skill is capable of reading and writing to the local filesystem to update project documentation and agent instructions in SKILL.md (Step 6).\n
  • Sanitization: There is no evidence of sanitization or validation for the content read from the repository before it is processed or stored.\n- [COMMAND_EXECUTION]: The skill uses git commands (git log, git diff, git status) to gather context about changes made during a coding session.\n- [EXTERNAL_DOWNLOADS]: The skill is authorized to fetch data from github.com according to its settings, which is a trusted and well-known service.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 7, 2026, 05:44 PM
Security Audit — agent-trust-hub — session-wrap-up-agent-skill