skills/xsavikx/okf-skills/okf-fs/Gen Agent Trust Hub

okf-fs

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill documents and ingests filesystem metadata, creating a vulnerability surface where malicious content in processed files could influence the AI agent.
  • Ingestion points: The skill reads metadata from .okf-metadata.yaml and content from OKF bundle Markdown files within main.go.
  • Boundary markers: The skill does not implement explicit delimiters or warnings to distinguish between file metadata and instructions for the AI agent.
  • Capability inventory: The skill has the ability to traverse directories, write new files to the output directory, and update local metadata files on the disk.
  • Sanitization: There is no evidence of sanitization or filtering of the description fields extracted from the filesystem before they are presented to the agent or written back to the disk.
  • [EXTERNAL_DOWNLOADS]: The skill instructions direct the user to download and install a Go binary from a remote repository.
  • Evidence: SKILL.md includes the command go install github.com/xSAVIKx/okf-skills/skills/okf-fs@v0.1.0. Note that this is a vendor-owned resource.
  • [COMMAND_EXECUTION]: The skill defines commands that execute a built Go binary to perform filesystem operations.
  • Evidence: The skill implements produce and ingest commands in main.go which perform directory traversal, file reads, and file writes.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 07:04 PM
Security Audit — agent-trust-hub — okf-fs