skills/b-open-io/prompts/remind/Gen Agent Trust Hub

remind

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: A detailed review of scripts/search.py, SKILL.md, and metadata found no evidence of malicious code, obfuscation, or unauthorized connectivity.
  • [DATA_EXFILTRATION]: Analyzed file access patterns in scripts/search.py. The script reads conversation logs from ~/.claude/projects/ and ~/.scribe/scribe.db. This access is necessary for the skill's intended purpose of recalling past sessions, and no network capabilities are present to exfiltrate this data.
  • [PROMPT_INJECTION]: Evaluated the surface for indirect prompt injection as the skill processes past conversation content. Ingestion points: Reads conversation text from local JSONL logs and SQLite database files. Boundary markers: None explicitly present in prompt interpolation. Capability inventory: Checked scripts/search.py; no subprocess calls, exec/eval, file-write operations, or network requests were found. Sanitization: The scribe_search function utilizes regex to strip HTML/XML tags and ANSI escape sequences from the retrieved content.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 03:57 AM