ClawHub
Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

Agent Phone Call

v1.0.2

Agent Phone Call gives every AI agent a phone number and the ability to get things done over the phone — finding contacts, making calls, and completing tasks...

⭐ 2· 23k·0 current·0 all-time
byFreeFly@mrsz
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
medium confidence
ℹ
Purpose & Capability
The declared purpose (giving agents a phone number and calling/answering functionality) matches the network endpoints and the included scripts (they call api.pollyreach.ai and agent.pollyreach.ai to register, send tasks, poll results, check balance, and handle inbound messages). Requesting read/write access to a PollyReach config file is consistent with storing a service token.
!
Instruction Scope
SKILL.md and the scripts disagree about the credentials filename (SKILL.md's registration step references ~/.config/PollyReach/credentials.json; all scripts read ~/.config/PollyReach/key.json). SKILL.md allows file read/write of key.json, but the text instructs saving to credentials.json — this mismatch can break behavior and may cause user confusion about where the token is stored. The scripts only contact pollyreach.ai endpoints and only read the single token file (no other files), which is expected, but the discrepancy in filename and an undocumented environment override (POLLYREACH_KEY_FILE) are scope/confusion risks.
✓
Install Mechanism
There is no installer that downloads arbitrary external code; the skill is instruction + included shell scripts (bundled in the skill). Dependencies are standard CLI tools (curl, jq, bc). No unusual network install steps or third‑party archive downloads were found.
ℹ
Credentials
The skill requests access to a local credentials file in the user's home (expected for storing the service token) and network access to pollyreach.ai (expected). However, the registry metadata claims no required binaries/env vars while SKILL.md lists curl/jq/bc — a small inconsistency. Also scripts support the POLLYREACH_KEY_FILE env var (to override file path) but this env var is not declared in the manifest. No unrelated secrets are requested.
✓
Persistence & Privilege
The skill is not marked always:true and does not request system‑wide privileges. It will read/write its own config file in the user's home directory, which is normal for storing tokens. It does not modify other skills or global agent settings.
What to consider before installing
This skill largely does what it says, but there are a few mismatches and trust considerations to address before installing: - Filename mismatch: the SKILL.md instructs saving the token to ~/.config/PollyReach/credentials.json but every script reads ~/.config/PollyReach/key.json. Decide which file will be used (or set POLLYREACH_KEY_FILE) and make sure the token is stored in the file the scripts actually read. - Verify the destination: all network calls go to api.pollyreach.ai and agent.pollyreach.ai. Confirm you trust PollyReach and their privacy/billing practices before giving the service an account and token. - Token safety: the skill stores a bearer token on disk. Use a dedicated/throwaway PollyReach account if possible, monitor usage/charges, and remove the token if you stop using the skill. - Environment override: scripts support POLLYREACH_KEY_FILE but this is undocumented in the manifest — if you prefer a different storage location, set that env var explicitly. - Test safely: try the skill in an isolated environment (non‑production account or VM) first to confirm behavior and to verify which credential filename is actually used. If you want, I can: (a) suggest a corrected SKILL.md and scripts alignment (unify to one filename), (b) produce a short checklist to safely onboard a PollyReach token, or (c) highlight exact lines to change so the skill is consistent.

Like a lobster shell, security has layers — review code before you run it.

latestvk971frmxf6a340wsvfkkrhzv4d842k9a

License

MIT-0
Free to use, modify, and redistribute. No attribution required.
Termshttps://spdx.org/licenses/MIT-0.html

Comments