OC SRE On-Demand Service

Provides expert SRE support for OpenClaw environments including health checks, connectivity fixes, execution stabilization, and security hardening guidance.

Security Scan

This is an instruction-only SRE engagement blueprint (no code, no installs, no credentials requested) that is internally consistent but very high-level and underspecified for real-world operations.

Purpose & Capability

The name and description claim SRE duties (health checks, connectivity fixes, hardening). The SKILL.md is a service/process blueprint that matches that purpose, but it does not request any credentials, tooling, or access methods that would normally be needed to perform those tasks—so it is coherent as a playbook but under-specified for doing live SRE work.

Instruction Scope

The runtime instructions are high-level process guidance (intake/diagnosis/remediation/verification) and include a data-handling note to sanitize external reports. They do not instruct the agent to read files, call external endpoints, or access environment variables. However, the guidance is open-ended and grants broad discretion to the agent about what 'evidence-first checks' and 'smallest safe fix' mean in practice.

Install Mechanism

No install spec and no code files are present; this is instruction-only so nothing is written to disk or fetched during install.

Credentials

The skill requests no environment variables, credentials, or config paths (proportionate for a playbook). Note: performing actual SRE actions would typically require credentials or access; the skill does not attempt to obtain them automatically.

Persistence & Privilege

always:false and user-invocable:true (normal). The skill does not request persistent presence or modify other skills; autonomous model invocation is allowed by default but is not itself a red flag here.

Assessment

This skill is a benign, high-level SRE playbook rather than an executable tool. Before using it in a production context: 1) Verify the publisher and provenance (source is unknown). 2) Do not hand over credentials or broad access tokens to the agent—real fixes should be performed by an authorized operator or in a constrained/test environment. 3) Require explicit, auditable consent before any remedial actions; prefer step-by-step human approval. 4) Request logs, sanitized outputs, and precise action proposals rather than allowing the agent to act autonomously. 5) If you plan to let the agent execute changes, only grant least-privilege, time-limited credentials and enable full auditing. If any SKILL.md revision adds commands, network calls, or credential access, reevaluate immediately.

Service scope

• Gateway and node health checks
• Pairing/connectivity incident response
• Execution-path stabilization
• Security and configuration optimization

Engagement model

1. Intake: issue summary and target environment.
2. Diagnosis: evidence-first checks and hypothesis.
3. Remediation: smallest safe fix first.
4. Verification: before/after proofs.
5. Handoff: actionable next steps.

Output contract

• findings
• actions taken
• verification
• residual risk
• recommended roadmap

Data handling

• external-facing reports must be fully sanitized.
• use placeholders for endpoints, infra, and identities.