Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
postgres-tool
v1.0.0PostgreSQL database management tool for querying databases, exporting results, inspecting schemas, and safely performing UPDATE/DELETE operations with automa...
⭐ 0· 70·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
Name/description, scripts, and docs align with a Postgres management utility (querying, export, schema inspection, safe UPDATE/DELETE with backups). There are no unexpected external-service credentials or unrelated binaries requested. However the package includes a config/db_config.json with plaintext credentials (127.0.0.1:5432, user postgres, password postgres) bundled into the repo — reasonable for examples but potentially dangerous if left in production deployments.
Instruction Scope
SKILL.md instructs the agent to invoke the packaged scripts directly and explicitly says the scripts will 'automatically search and load' db_config.json without user-specified path. The implementation searches the current working directory first, then skill/config, then skill root — meaning the skill may read a db_config.json anywhere the agent runs (potentially picking up sensitive local configs). The instructions do not ask the agent to access unrelated system files or external endpoints, but the implicit auto-load behavior increases the chance of unintended credential use/exposure.
Install Mechanism
No install spec in registry; installation is manual (pip install or offline wheels). All dependency instructions use standard pip and local wheel directories; there are no downloads from unknown URLs or extract-from-remote steps in the bundle. The offline install scripts are local and expected for an internal/offline workflow.
Credentials
The skill declares no required env vars or primary credential, but the code expects a db_config.json file containing DB credentials. Shipping a plaintext config with default admin-like credentials inside the package (config/db_config.json) is disproportionate and risky. The design relies on file-based credentials instead of declared environment variables, and the auto-search across working directories increases the likelihood of loading unintended/sensitive credentials. Requiring DB credentials is reasonable for the tool's purpose, but how they are provided (and the bundled example credentials) should be reviewed.
Persistence & Privilege
The skill does not request always:true, does not modify other skills, and will only write backups into its own SKILL_ROOT/backups directory. It can run autonomously by default (normal), but it does not request elevated system-wide privileges.
Scan Findings in Context
[hardcoded-credentials] unexpected: The bundle contains config/db_config.json with plaintext credentials (host 127.0.0.1, user 'postgres', password 'postgres'). Example configs are common, but bundling working credentials in the package can lead to accidental use or leakage and is not necessary for the skill's functionality.
What to consider before installing
This skill appears to do what it claims, but review the following before installing: 1) Remove or replace the bundled config/db_config.json — do not leave plaintext credentials in the package or deploy it as-is to production. 2) Be aware the scripts auto-load db_config.json from your current working directory first — ensure no sensitive db_config.json exists in any directories where the agent may run. 3) Inspect scripts/postgres_tool.py to confirm backup locations and restore behavior meet your policies (it writes backups under the skill directory). 4) Only install dependencies from trusted sources (the repo expects local wheel files or pip installs). 5) Run the skill in an isolated/test environment first and confirm interactive confirmation behavior (input prompts) or adjust automation flags (like --force) carefully. If you want a higher assurance, ask the author to remove default credentials and require explicit config or environment variables for DB access.Like a lobster shell, security has layers — review code before you run it.
latestvk978wmfeecjwbjv5spn4fk5re583ez78
License
MIT-0
Free to use, modify, and redistribute. No attribution required.