skills/gregbarbosa/skills/handoff/Gen Agent Trust Hub

handoff

Pass

Audited by Gen Agent Trust Hub on Oct 8, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes shell commands including git rev-parse, git status, git branch, and mkdir to determine the environment context and create the handoff directory structure.
  • [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read, verify, and execute tasks defined in external handoff files (stored in ${HANDOFF_DIR:-$HOME/.claude/handoffs}) during session resumption. This ingestion of file content as instructions for task creation (TaskCreate) constitutes an indirect prompt injection surface, although it is the core intended functionality of the skill.
  • [DATA_EXPOSURE]: The skill captures session state, including work in progress and hard rules, to local files. It mitigates risk by explicitly instructing the agent that the file 'carries no secrets or tokens.'
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 8, 2026, 09:43 PM