Back to skill
Skillv1.0.0
ClawScan security
Reactive Resume · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignMar 15, 2026, 9:46 AM
- Verdict
- benign
- Confidence
- medium
- Model
- gpt-5-mini
- Summary
- The skill is a coherent developer guide + helper scripts for Reactive Resume; the included scripts can drop/reset a development database and rely on Docker and env vars, so use only in a development environment and inspect .env and scripts before running.
- Guidance
- This skill is a development guide and includes helper scripts; it is not trying to hide malicious behavior, but you should be careful before running anything: 1) Inspect .env/.env.example — the docs expect DATABASE_URL, PRINTER_ENDPOINT, and other secrets; never point DATABASE_URL at a production DB. 2) The scripts include db-reset.py which will drop all tables via psql — run it only against a dev database and read the script to understand its prompts. 3) db-reset.py uses subprocess with shell=True and constructs shell commands from DATABASE_URL; avoid running it on untrusted input. 4) create-template.py writes files into your repo; review paths before running. 5) The docs recommend starting dockerd with sudo and running docker compose — these are normal for self-hosting but require root and network access. If you plan to let an agent invoke this skill autonomously, be aware it may instruct the agent to run system commands that can modify your environment; prefer manual invocation and confirm environment variables and target systems first.
Review Dimensions
- Purpose & Capability
- noteName/description match the content: this is a development/self-hosting guide for Reactive Resume. The included files (template boilerplate, template-creator script, DB reset script, and extensive docs) are appropriate for that purpose. Minor inconsistency: skill metadata declares no required environment variables, but the documentation and scripts clearly require several environment variables to run (DATABASE_URL, PRINTER_ENDPOINT, BETTER_AUTH_SECRET, optional cloud/storage/API keys).
- Instruction Scope
- noteSKILL.md stays within the stated scope (local dev, templates, deployment). It instructs running Docker, docker-compose, copying .env, starting the dev server, and performing DB migrations. It also references scripts that will read .env and run psql to drop tables — destructive but consistent with a 'dev reset' tool. Nothing in SKILL.md instructs exfiltration or accessing unrelated system resources, but it does instruct running system-level commands (sudo dockerd, docker compose, pnpm, psql).
- Install Mechanism
- okNo install spec — instruction-only with bundled helper scripts and docs. That's low-risk for arbitrary code download. The provided files are typical project artifacts; there are no external arbitrary download URLs or extraction steps in the install metadata.
- Credentials
- concernThe skill metadata lists no required env vars, yet the documentation and scripts depend on many environment variables (DATABASE_URL, PRINTER_ENDPOINT/APP_URL, BETTER_AUTH_SECRET, STORAGE_* and optional AI keys). The db-reset script reads .env or env vars and will act on the database referenced by DATABASE_URL. This mismatch between declared requirements and actual runtime needs could cause surprising behavior if run without awareness.
- Persistence & Privilege
- okSkill is not always-enabled and does not request persistent platform-wide privileges. It does not modify other skills or system-wide agent config. The biggest operational privilege is the ability to run user-level system commands when a developer follows the guide (expected for a dev tool).