AI Security Review
scanned 2h ago · by lpm-firewall-aiReview flagged AI-agent configuration or capability changes. This remains warn-only unless evidence shows foreign-agent hijack through preinstall/install/postinstall, hidden persistence, exfiltration, remote code execution, or other concrete malicious behavior.
Decision evidence
public snapshot- `bin/cli.mjs` explicit `install` copies a skill into `~/.claude`, `~/.codex`, or `~/.hermes`.
- For Claude targets, `bin/cli.mjs` deletes and replaces package-owned `hyv-*.md` agent files in `.claude/agents`.
- `install.sh` provides the same explicit agent-extension installation path.
- `veo-generate.py` dynamically loads a user-supplied provider plugin path.
- `package.json` has only `prepublishOnly`; no install-time lifecycle hook.
- Agent/config writes require an explicit CLI or shell `install` invocation.
- Agent cleanup is scoped to the package-owned `hyv-*.md` filename prefix.
- Network use is feature-aligned: CC-image APIs and ElevenLabs TTS.
- `gen-voiceover.mjs` requires `--yes` before sending API-key-authenticated TTS requests.
- The bundled `.xlsx` is a standard XML workbook with no macro, OLE, or external-link entries.
Source & flagged code
4 flagged · loading sourcePackage ships high-entropy non-source blobs.
skills/hearyourvoice/references/examples/chado-NG-shotlist.xlsxView on unpkgPackage ships compressed or archive-like blobs.
skills/hearyourvoice/references/examples/chado-NG-shotlist.xlsxView on unpkgPackage ships a nested archive or MCP bundle that was inventoried but not recursively analyzed.
skills/hearyourvoice/references/examples/chado-NG-shotlist.xlsxView on unpkg