OC Full Ops Audit Recipe

End-to-end OpenClaw audit and remediation recipe for gateway, channels, nodes, security, and memory sync.

Security Scan

The skill is a high-level, instruction-only audit recipe whose instructions and requirements are internally consistent, but it is vague about making changes and about where findings are stored — review memory and approval policies before use.

Purpose & Capability

Name and description describe an end-to-end audit recipe and the SKILL.md contains a concise, matching workflow (baseline, classify, apply fixes, re-check, document). There are no extra binaries, env vars, or installs that would be unexpected for this purpose.

Instruction Scope

The instructions are high-level and remain within the scope of an audit recipe. However, two steps need clarification: (1) 'Apply fixes in risk order with rollback points' implies performing potentially disruptive actions but provides no guardrails, required credentials, or explicit approval step; (2) 'Write outcomes to daily memory + shared memory' instructs persistence of results but does not define what memory stores are, retention, or redaction rules. Both are scope-expanding details that should be explicit.

Install Mechanism

This is an instruction-only skill with no install spec and no code files — lowest-risk install profile. Nothing will be written or fetched by an installer during skill installation.

Credentials

The skill declares no required environment variables or credentials, which is consistent with a read-only audit. However, because it instructs applying fixes (which typically require elevated credentials) and writing results to shared memory, the absence of declared credential requirements is a gap — any implementation that actually performs fixes will need credentials that aren't described here.

Persistence & Privilege

always:false and user-invocable:true are appropriate. Still, the SKILL.md explicitly directs writing outcomes to persistent 'daily memory' and 'shared memory', which could cause retention of sensitive audit findings. Combine that with the platform's default autonomous invocation and you should confirm memory access controls, retention, and whether the skill will ask for explicit approval before making changes.

Assessment

This skill is a coherent, high-level audit recipe and appears benign, but it is intentionally vague about two important things: making changes and where results are stored. Before installing or running it, confirm: (1) where 'daily memory' and 'shared memory' live, who can read them, how long entries are retained, and whether outputs will be redacted to avoid leaking secrets; (2) that the agent will require and request explicit, itemized approval before performing any disruptive fixes, and that any credentials needed for remediation are managed and consented to separately; (3) that rollback points and verification steps are implemented so changes can be safely reverted. If you want a safer baseline, restrict this skill to read-only checks until remediation procedures, credential handling, and memory/retention policies are explicitly documented.

Goal

Provide a repeatable audit workflow that ends with verified fixes and documented memory updates.

Audit flow

1. Baseline: status/health/gateway/security/nodes.
2. Classify findings: critical, warning, info.
3. Apply fixes in risk order with rollback points.
4. Re-run checks and compare deltas.
5. Write outcomes to daily memory + shared memory.

Deliverables

• Audit summary table
• Fix actions and verification
• Residual risks and priority plan

Usage notes

• Prefer read-only checks first.
• Group disruptive changes behind explicit approval.
• Keep outputs concise and operational.