Back to skill
Skillv1.1.1
ClawScan security
Community Outreach · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignFeb 21, 2026, 10:06 PM
- Verdict
- benign
- Confidence
- high
- Model
- gpt-5-mini
- Summary
- The skill's inputs, tools, and instructions line up with its stated purpose (finding community threads and drafting value-first responses); it requests no credentials, has no install steps, and is instruction-only.
- Guidance
- This skill appears coherent and low-risk: it only searches public communities and drafts outreach, and it requests no credentials or installs. Before installing, decide what file and web-access permissions to grant: the skill can read project docs (PRD/README) and will write docs/outreach-plan.md, so remove or sanitize any sensitive data in those files if present. If you don't want the agent to write files automatically, restrict or deny the Write permission or require explicit user approval. Review generated drafts before posting anywhere and ensure you follow each community's rules (and the skill already instructs disclosure and no vote-manipulation). If the platform exposes MCP tools (project_info, kb_search), consider whether you want the skill to use them and confirm their data sources.
Review Dimensions
- Purpose & Capability
- okName/description (find community threads and draft responses) align with the allowed tools and runtime instructions: the skill uses web searches and project files to locate threads and create drafts. It does request Read/Grep/Glob/Write capabilities so it can read project docs and write outreach-plan.md — which is consistent with the stated purpose, but review filesystem access scope if you want to limit what it can read.
- Instruction Scope
- okSKILL.md instructs only web searches, parsing project README/PRD, selecting threads, drafting responses, and writing a checklist/plan. It includes community-safety rules (disclose you're the developer, no vote manipulation). It does not instruct reading unrelated credentials or exfiltrating data. Note: it will write docs/outreach-plan.md and expects to read project docs; ensure those files don't contain sensitive data you don't want the skill to access.
- Install Mechanism
- okInstruction-only skill with no install spec and no code files — nothing is downloaded or written to disk beyond the planned output file. This is the lowest-risk install model.
- Credentials
- okThe skill requires no environment variables, no credentials, and no config paths. Its tooling (WebSearch/WebFetch and optional MCP search/project_info) is appropriate for web/community discovery; no secrets are requested.
- Persistence & Privilege
- notealways:false (normal). The skill is allowed to run autonomously by default (platform default) and to write a file (docs/outreach-plan.md). That writing behavior is coherent with its purpose, but if you do not want automatic file writes or autonomous runs, review and tighten tool permissions or require explicit invocation before allowing writes or web access.