Best fit
Reviewing selected stable OpenClaw configuration and instruction files for risky declarations.
// agent security
OpenClaw Guardrail Linter
Review selected OpenClaw configuration for risky declarations, with explicit reporting when inputs cannot be assessed.
Format: Agent skill, not a browser app.
Requirements: Python 3.10+ and json5==0.15.0. Install requirements.txt during approved setup; dependency installation needs package-source access. Scans are offline, with no API key or live OpenClaw connection.
You get: Rule findings and coverage status, using input ordinals and source hashes rather than raw values or absolute paths. Source line mapping is unavailable (line 0).
Package 1.1.0. Marketplace checkout may convert currency. Host/model charges are separate.
Is this the right fit?
It returns
Rule findings and coverage status, using input ordinals and source hashes rather than raw values or absolute paths. Source line mapping is unavailable (line 0).
Limits
Not a complete host-schema validator, vulnerability scanner or effective-permissions resolver. Includes are not followed. Other policy layers can constrain declarations; allowed shell execution can write despite file-tool denials.
Your first use
Extract the complete openclaw-guardrail-linter folder. Install the pinned dependency into an approved environment, then run the clean fixture below and the bundled risky fixture before selecting your own stable configuration.
python -m pip install -r requirements.txt
python scripts/scan_openclaw_guardrails.py fixtures/clean-sampleBack up customized files before updating. Automation must handle exit 2, not just 0 and 1. File checks do not promise resistance to adversarial concurrent replacement. Use native host policy/audit evidence to assess actual execution.
How to interpret the result
Exit 0: No findings in assessed inputs; review coverage and limitations. Exit 1: Findings need review. Exit 2: Assessment incomplete; do not treat as clean. None of these results is a security certification.
What changed in 1.1.0
Parsed JSON5 replaces regex-only configuration interpretation. Version 1.1.0 corrects file-mutation and neighboring-SecretRef findings, rejects duplicate keys and invalid UTF-8, and reports incomplete assessment for unsupported includes, malformed or missing inputs, linked/reparse paths and oversized files.
- Reads
- Selected local JSON5 configuration and instruction files. Credential/state paths are excluded without reading their contents.
- Version
- Package 1.1.0; Agensi platform revision 1.1
- Licence
- Review the package licence and seller terms at checkout; this update does not change them.
Questions
What does OpenClaw Guardrail Linter do?
Review selected OpenClaw configuration for risky declarations, with explicit reporting when inputs cannot be assessed.
What does it need to run?
Python 3.10+ and json5==0.15.0. Install requirements.txt during approved setup; dependency installation needs package-source access. Scans are offline, with no API key or live OpenClaw connection.
Does this prove my agent is secure?
Not a complete host-schema validator, vulnerability scanner or effective-permissions resolver. Includes are not followed. Other policy layers can constrain declarations; allowed shell execution can write despite file-tool denials.
Companion work
GuardrailDoctor covers a different review surface. Skill Injection Scanner reviews a skill package before installation.
Check your agent setup
Review tested hosts, versions and workflow limits. Evidence for another skill does not establish this product’s compatibility; keep the requirements and limitations above in view.