skills/getpaseo/paseo/paseo-handoff/Gen Agent Trust Hub

paseo-handoff

Pass

Audited by Gen Agent Trust Hub on Mar 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses git branch --show-current to identify the base branch for worktrees and paseo run to launch a new agent instance. These commands are essential for the skill's intended purpose of task delegation and orchestration.
  • [PROMPT_INJECTION]: This skill presents a surface for indirect prompt injection as it aggregates current conversation context and file content into a prompt for a downstream agent. 1. Ingestion points: Gathers data from $ARGUMENTS and the existing workspace/session context within SKILL.md. 2. Boundary markers: Utilizes structured Markdown headers (e.g., '## Task', '## Context') to delimit sections of the handoff prompt, though it lacks explicit instructions for the receiving agent to disregard potential commands within the provided context. 3. Capability inventory: Executes local CLI tools git and paseo as specified in the bash snippets. 4. Sanitization: Aggregates and passes context data into the handoff template without performing specific sanitization or filtering of potential injection payloads.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 15, 2026, 02:09 PM