obsidian-tag-cleanup-flow
Pass
Audited by Gen Agent Trust Hub on May 19, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: No attempts to override agent behavior or bypass safety guidelines were detected. The instructions follow a structured, stage-based workflow logic.
- [DATA_EXFILTRATION]: The skill operates entirely on local Obsidian vault data (.md files). There are no network operations (curl, wget, etc.) or hardcoded credentials detected.
- [REMOTE_CODE_EXECUTION]: No remote script downloads or unverified package installations are present. The skill functions by calling other internal agent skills.
- [OBFUSCATION]: The content is clear-text markdown with no Base64, zero-width characters, or other encoding techniques used to hide malicious intent.
- [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface as it ingests untrusted data (user notes) and has file-write capabilities (modifying tags/metadata).
- Ingestion points: Reads tag lists and note content from the Obsidian vault.
- Boundary markers: None explicitly defined in this orchestrator level.
- Capability inventory: File write operations (tag renaming, merging, and retirement) across Obsidian notes.
- Sanitization: Relies on the underlying specialty skills (e.g., obsidian-tag-hygiene) for processing.
- [COMMAND_EXECUTION]: No shell commands or system-level process executions are initiated by this skill.
Audit Metadata