Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed PowPow command-line helper whose login, posting, deletion, upload, and verification behavior matches its stated purpose and includes confirmation gates for writes.
Install only if you trust PowPow with your account session and understand that confirmed actions can publish public posts, upload images, create public digital humans that cost badges, or irreversibly delete your own posts. Use the dry-run previews first, add --confirm only after reviewing the output, and run logout on shared machines.
The skill contains standalone read/verification behavior not reflected in the concise declared purpose. While not inherently malicious, undocumented verification endpoints and read-side API calls still matter in a security review because they can leak identifiers, confirm resource existence, or create unanticipated audit traffic.
The skill contains standalone read/verification behavior not reflected in the concise declared purpose. While not inherently malicious, undocumented verification endpoints and read-side API calls still matter in a security review because they can leak identifiers, confirm resource existence, or create unanticipated audit traffic.
The skill contains standalone read/verification behavior not reflected in the concise declared purpose. While not inherently malicious, undocumented verification endpoints and read-side API calls still matter in a security review because they can leak identifiers, confirm resource existence, or create unanticipated audit traffic.
The skill contains standalone read/verification behavior not reflected in the concise declared purpose. While not inherently malicious, undocumented verification endpoints and read-side API calls still matter in a security review because they can leak identifiers, confirm resource existence, or create unanticipated audit traffic.
The skill contains standalone read/verification behavior not reflected in the concise declared purpose. While not inherently malicious, undocumented verification endpoints and read-side API calls still matter in a security review because they can leak identifiers, confirm resource existence, or create unanticipated audit traffic.
The skill contains standalone read/verification behavior not reflected in the concise declared purpose. While not inherently malicious, undocumented verification endpoints and read-side API calls still matter in a security review because they can leak identifiers, confirm resource existence, or create unanticipated audit traffic.
The skill contains standalone read/verification behavior not reflected in the concise declared purpose. While not inherently malicious, undocumented verification endpoints and read-side API calls still matter in a security review because they can leak identifiers, confirm resource existence, or create unanticipated audit traffic.
The skill contains standalone read/verification behavior not reflected in the concise declared purpose. While not inherently malicious, undocumented verification endpoints and read-side API calls still matter in a security review because they can leak identifiers, confirm resource existence, or create unanticipated audit traffic.
The skill contains standalone read/verification behavior not reflected in the concise declared purpose. While not inherently malicious, undocumented verification endpoints and read-side API calls still matter in a security review because they can leak identifiers, confirm resource existence, or create unanticipated audit traffic.
The skill contains standalone read/verification behavior not reflected in the concise declared purpose. While not inherently malicious, undocumented verification endpoints and read-side API calls still matter in a security review because they can leak identifiers, confirm resource existence, or create unanticipated audit traffic.
The skill contains standalone read/verification behavior not reflected in the concise declared purpose. While not inherently malicious, undocumented verification endpoints and read-side API calls still matter in a security review because they can leak identifiers, confirm resource existence, or create unanticipated audit traffic.
The skill contains standalone read/verification behavior not reflected in the concise declared purpose. While not inherently malicious, undocumented verification endpoints and read-side API calls still matter in a security review because they can leak identifiers, confirm resource existence, or create unanticipated audit traffic.
Referenced artifact was not completely inspected
node scripts/geocode.js "<地名>" [--city <城市>] [--limit N] [--json]
Referenced artifact was not completely inspected
node scripts/list-digital-humans.js [关键词] [--all] [--limit N] [--json]
Referenced artifact was not completely inspected
node scripts/match-digital-human.js <topic> [--limit N] [--json]
Referenced artifact was not completely inspected
node scripts/search-image.js <query | unsplash-photo-page-url>
Referenced artifact was not completely inspected
node scripts/verify.js <post-id>
Referenced artifact was not completely inspected
node scripts/compose.js --text-file <path> [--dh <id|name:<名>|auto>] [--topic "<topic>"] \
Referenced artifact was not completely inspected
node scripts/create-digital-human.js --name ... --desc ... --lng x --lat y --avatar ... --confirm
Referenced artifact was not completely inspected
node scripts/delete-post.js <post-id> # 拉取帖子摘要，exit 3
Referenced artifact was not completely inspected
node scripts/delete-post.js <post-id> # 拉取帖子摘要，exit 3
Referenced artifact was not completely inspected
node scripts/upload-image.js <local-image-path> # 只打印上传计划，exit 3
Referenced artifact was not completely inspected
node scripts/upload-image.js <local-image-path> # 只打印上传计划，exit 3
Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).
/**
* Delete one of your own posts (DELETE /api/posts/[id], JWT-scoped).
* Useful for cleaning up test posts. You can only delete your own posts.
*
* CONFIRMATION GATE: without --confirm this only FETCHES the post and prints
The skill declares network- and environment-using behavior but does not specify any explicit tool scope such as allowed-tools or permissions. In an agent setting, that omission weakens least-privilege guarantees and can let downstream orchestrators grant broader access than intended, especially since the skill handles authentication, local session state, and remote API interaction.
Detected: suspicious.env_credential_access, suspicious.potential_exfiltration