Back to skill
Skillv3.0.3
ClawScan security
Brand Namer · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignMar 19, 2026, 1:02 AM
- Verdict
- Benign
- Confidence
- high
- Model
- gpt-5-mini
- Summary
- The skill does what it says (generate names, check domains, analyze and save candidates); scripts are bundled and perform local checks and DNS/HTTP probes and store results under ~/.brand-namer, with only minor documentation inconsistencies.
- Guidance
- This skill appears to be what it claims: name generation, local analysis, and domain availability checks. Before installing, note: (1) it will create ~/.brand-namer and write saved.txt and history.log there; (2) domain checks perform outbound network probes (dig or HTTP requests via curl) to domains built from your inputs — if you need to avoid any outbound network traffic, do not run the check commands; (3) documentation and metadata mismatch (SKILL.md mentions shortlist.json and 'dig' as required while the scripts use saved.txt and expect dig/curl) — verify the scripts match your expectations. If you are comfortable with those behaviors, the skill is coherent and safe to use. If unsure, run the scripts in a sandboxed environment or inspect/modify them before use.
Review Dimensions
- Purpose & Capability
- noteName/description match the provided scripts: generators, domain checks, analysis, and shortlist management. Minor inconsistencies: SKILL.md claims saved data is 'shortlist.json' while the actual script uses saved.txt (SAVED_FILE="$DATA_DIR/saved.txt"). Also SKILL.md documents 'dig' as required but the registry metadata lists no required binaries.
- Instruction Scope
- noteRuntime instructions tell the agent to run bundled scripts (scripts/script.sh). The scripts read/write only to a dedicated data directory (~/.brand-namer) and perform DNS/HTTP checks (dig or curl) against domains derived from user input. They do not access other user files or external hidden endpoints. The bundled Python variant (scripts/brand.sh) uses curl for HTTP probing of many TLDs (different behavior than the bash script), so domain checks may trigger outbound HTTP requests to arbitrary domains constructed from user input.
- Install Mechanism
- okNo install spec or external downloads; this is instruction-only with bundled scripts. Nothing is pulled from arbitrary URLs and no archive extraction occurs.
- Credentials
- noteThe skill requests no credentials or sensitive environment variables. It does create and write files under $HOME/.brand-namer. The SKILL.md expects the 'dig' utility and bash 4.0+, but the package metadata did not declare required binaries—this mismatch should be fixed but is not a sign of exfiltration.
- Persistence & Privilege
- okalways:false and user-invocable. The skill writes only to its own data directory and a history log there; it does not modify other skills or system-wide agent settings.