startup
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes standard Git commands and shell utilities to determine the current project state.
- Evidence includes the use of
git status --short,git fetch origin, andgit rev-list --count "HEAD..@{u}"to check synchronization with remote repositories. - It retrieves the local system's identity using
$env:COMPUTERNAME(PowerShell) orhostname(Mac/Linux) to detect if work was recently performed on a different machine. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and summarize external project files, creating a potential surface for indirect prompt injection.
- Ingestion points: The skill reads
AGENTS.md,handoff.md, and potentially Obsidian notes files. - Boundary markers: The instructions do not specify explicit delimiters for the ingested content, though they instruct the agent to provide a "structured summary" rather than a full output.
- Capability inventory: The agent has access to file system reads, shell command execution (hostname), and Git operations.
- Sanitization: No specific sanitization or escaping of the ingested file content is mentioned.
Audit Metadata