upward-networking
Pass
Audited by Gen Agent Trust Hub on Aug 4, 2026
Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
- [NO_CODE]: The skill consists entirely of natural language instructions in Markdown and does not contain any executable scripts, binaries, or automated shell commands.\n- [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it ingests user-provided descriptions of interactions and networking targets to generate formatted output. Although the skill has instructions to require user confirmation and vault path specification before saving notes, it lacks explicit sanitization requirements for the processed data.\n
- Ingestion points: User-provided descriptions of networking targets, current relationship status, and interaction goals processed in SKILL.md.\n
- Boundary markers: The skill includes specific logical constraints requiring the agent to obtain user confirmation and a valid vault path before performing any file-write operations.\n
- Capability inventory: Filesystem write capability to create or append Obsidian Markdown files as defined in the Obsidian Mode section.\n
- Sanitization: No explicit instructions for sanitizing or escaping user-supplied content are included before the data is persisted to the filesystem.
Audit Metadata