Back to skill
Skillv1.0.0
ClawScan security
pidgesms · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignFeb 22, 2026, 10:31 PM
- Verdict
- benign
- Confidence
- high
- Model
- gpt-5-mini
- Summary
- The skill's requirements and runtime instructions match its stated purpose (using the pidge CLI to send/read SMS via an Android SMS Gateway); nothing requests unrelated credentials or surprising install sources, though you should trust the upstream pidge project before installing.
- Guidance
- This skill appears coherent and implements exactly what it claims: a thin wrapper around the pidge CLI to interact with an Android SMS Gateway. Before installing, verify you trust the upstream pidge project (SKILL.md suggests installing via `go install github.com/typhonius/pidge@latest` which fetches code from GitHub). Secure your phone and gateway (run the Android SMS Gateway only on trusted networks), inspect ~/.config/pidge/config.toml for any credentials or webhook endpoints, and test with a controlled phone number. If you want to prevent accidental or automated sends, consider disabling autonomous skill invocation for this skill (or require explicit user approval) so the agent cannot send SMS without your confirmation.
Review Dimensions
- Purpose & Capability
- okName/description (SMS via pidge) align with the declared runtime requirements: the skill only requires the pidge binary and the pidge config (~/.config/pidge/config.toml). There are no unrelated environment variables, binaries, or config paths requested.
- Instruction Scope
- okSKILL.md instructs only on running the pidge CLI (send, inbox, status, ack/unack, health) and references pidge's config file. It does not instruct reading other files, scanning the system, or posting SMS content to external endpoints unrelated to the Android SMS Gateway.
- Install Mechanism
- noteThe registry lists no formal install spec, but the SKILL.md metadata suggests `go install github.com/typhonius/pidge@latest`. That will fetch and build code from the public GitHub repo (moderate trust requirement). This is not a private/paste URL or arbitrary binary download, but it does pull external code to disk — review the upstream project before installing.
- Credentials
- okNo environment variables or credentials are requested. The only config path referenced is pidge's config (~/.config/pidge/config.toml), which is appropriate for the stated purpose.
- Persistence & Privilege
- noteThe skill does not request permanent/always-on inclusion (always:false) and does not modify other skills. It is callable by the model (default behavior) — because the skill can send real SMS, consider the implications of autonomous invocation (the platform default) if you need to prevent the agent from sending messages without explicit approval.