Picoclaw Security Posture Review
Review Picoclaw security posture in read-only mode and give operators safe remediation guidance.
The code explicitly states that it performs local, read-only posture review, has no runtime network egress, runs on demand, and does not auto-remediate; installation guidance includes signature and checksum verification. Points are deducted because profile inputs, sensitive-data flows, privilege isolation, rollback/recovery, and operator-confirmation boundaries are not fully specified, while installation depends on external network access.
The core implementation is small, deterministic in its formatted output, and has one unit test covering the main finding categories. Points are deducted because coverage is narrow, with no evidence for nulls, wrong types, malformed JSON, unknown fields, or a real Picoclaw environment; CLI file and JSON errors are surfaced as raw exceptions with limited diagnostic handling.
SKILL.md, README, and skill.json consistently target Picoclaw, local posture profiles, read-only findings, and explicit trigger phrases, while excluding other platforms. Points are deducted for no input schema, output examples, detailed non-fit boundaries, or Chinese-language support; standalone release verification depends on GitHub and other overseas services, although the core local check does not require network access.
The package includes version metadata, AGPL-3.0-or-later licensing, an SBOM, installation and quickstart guidance, validation commands, and a changelog. Points are deducted because README still labels the package v0.0.1 while metadata is v0.0.5, maintenance ownership and update path are unclear, and troubleshooting and compatibility policy are thin; publisher identity is not verified by the curated registry.
The implementation can produce severity-ranked findings, summary counts, and remediation guidance as directly consumable JSON. Points are deducted because the rules depend on an undocumented profile structure, cover only a limited set of boolean markers, and do not establish complete or accurate posture auditing for real configurations; the static-review ceiling limits this dimension to 7.
The repository contains the implementation, a corresponding unit test, changelog entries, and release-artifact verification instructions, providing limited auditability. Points are deducted because no tests were executed in this review, the committed test covers only one constructed input, and there is no third-party execution evidence or corroborating source; the static-review ceiling limits this dimension to 5.
- Do not describe this as a complete security suite; it is a Picoclaw-only local read-only rule checker.
- Confirm the actual posture-profile schema and field semantics before relying on results, or false negatives and false positives may occur.
- The verification installation flow accesses GitHub and invokes external tools including curl, openssl, and jq; validate the pinned public-key fingerprint in a controlled environment.
- README, SKILL.md, and skill.json contain inconsistent version information and should be aligned before release.
- The CLI lacks friendly failure handling for missing files, permission errors, and invalid JSON.
What does this skill do, and when should you use it?
This is a standalone Picoclaw security review skill in the ClawSec repository, focused on local, read-only posture analysis. It reads an existing Picoclaw security profile and checks for exposed Web UI, authentication weaknesses, unrestricted workspace or tooling, unsigned verification, MCP trust-boundary concerns, scheduler persistence, plaintext secret markers, and multi-channel authentication issues. It does not change configuration; it reports findings for the operator to address. It fits teams that want a focused Picoclaw deployment review without automatic remediation.
Runs node scripts/self_pen_test.mjs with a --profile argument pointing to an existing Picoclaw posture profile. It checks public Web UI exposure, disabled UI authentication, unrestricted workspace or tooling, unsigned verification mode, MCP trust-boundary review needs, scheduler persistence, plaintext secret markers, and multi-channel authentication. The analysis is local and read-only; findings and remediation guidance are provided for the operator rather than applied automatically. For standalone releases, the documented verification flow uses curl, OpenSSL, jq, and hashing utilities to validate the signed release manifest, archive, SKILL.md, and skill.json.
- A Picoclaw operator wants a pre-deployment review of an existing local security posture profile.
- A security reviewer needs to identify exposed Web UI, missing authentication, or overly broad tooling permissions.
- A team needs to review MCP trust boundaries and scheduler persistence without allowing an audit tool to change production settings.
- An administrator wants to verify a standalone release package before installing or extracting it.
What are this skill's strengths and limitations?
- Clearly scoped to Picoclaw local security posture review.
- Read-only behavior reduces the risk of automatic production configuration changes.
- Covers Web UI, authentication, tooling, MCP, scheduler, and secret-marker concerns.
- Documents signed release-manifest and file-integrity verification.
- It analyzes an existing posture profile and does not document how to discover or generate one.
- It does not automatically remediate findings.
- Runtime use requires Node.js; standalone release verification additionally requires curl, OpenSSL, jq, and hashing tools.
- The supplied material does not establish support for platforms other than Picoclaw or provide pricing information.
How do you install this skill?
Install with the Vercel Skills CLI: npx skills add prompt-security/clawsec --skill picoclaw-self-pen-testing -a openclaw -y. For standalone installs, follow the documented release-manifest verification procedure first; install or extract the archive only after verification succeeds. The repository also documents npx clawhub@latest install clawsec-suite, but that installs the broader suite rather than providing a dedicated installation command for this skill.
How do you use this skill?
Provide an existing Picoclaw posture profile, then run: node scripts/self_pen_test.mjs --profile ~/.picoclaw/security/clawsec/current-profile.json. The supplied material does not document a trigger prompt, output schema, or exact remediation commands. Maintainers can validate the package with python utils/validate_skill.py skills/picoclaw-self-pen-testing and node skills/picoclaw-self-pen-testing/test/self_pen_test.test.mjs.
How does this skill compare with similar options?
Compared with the repository's picoclaw-security-guardian, this skill focuses on agent self-pen-testing and read-only posture review. The README describes the guardian as covering Picoclaw advisories, configuration drift, and supply-chain checks. Choose the guardian when those capabilities are required; they should not be attributed to this skill.