MaxClaw Helper

Usage guidance and runtime diagnostics for MaxClaw / OpenClaw. Activate when: (1) User asks how to use a feature — e.g. 'how to enable Feishu long polling',...

Security Scan

An instruction-only troubleshooting/usage helper for MaxClaw/OpenClaw that is internally consistent with its stated purpose: it cites trusted docs, contains runtime guidance, requests no installs or credentials, and does not attempt out-of-scope access.

Purpose & Capability

The skill is a usage and diagnostics helper for MaxClaw/OpenClaw and its instructions focus on querying documentation (OpenClaw docs and a MaxClaw Feishu Wiki), classifying runtime failures, and escalating to community if needed. The referenced sources (OpenClaw docs and MaxClaw Feishu Wiki) are appropriate for the stated purpose.

Instruction Scope

SKILL.md contains explicit flows for doc lookup, failure classification, and escalation. It tells the agent to query documentation and ask the user for config / error messages when needed. It references reading config like openclaw.json only to decide to stop or do safety checks (i.e., it does not instruct blanket reading/exfiltration). One small caveat: the Feishu Wiki flow shows steps that involve obtaining an obj_token and reading a document via feishu-wiki/feishu-doc actions — this behavior is expected for a MaxClaw helper but does implicitly rely on platform-provided access to Feishu; the skill itself does not instruct collecting unrelated secrets or sending data to unexpected endpoints.

Install Mechanism

No install spec and no code files are present (instruction-only). This minimizes disk writes and arbitrary code execution risk.

Credentials

The skill declares no required env vars, credentials, or config paths, which is proportional. However, the trusted-sources instructions include steps for accessing a Feishu Wiki (obj_token/doc_token). Those steps imply the agent/platform must have some Feishu access mechanism; the skill does not request explicit credentials itself. This is likely acceptable if the platform exposes feishu-wiki/feishu-doc capabilities, but you should verify how Feishu access is provisioned and whether any tokens could be used.

Persistence & Privilege

always:false and no install behavior means the skill has no special persistent presence or elevated privileges. It does not modify other skills or system-wide settings in its instructions.

Assessment

This skill appears coherent and low-risk because it is instruction-only, asks for no credentials, and points to official docs. Before installing, consider: (1) Confirm how your agent/platform provides access to the Feishu Wiki (feishu-wiki/feishu-doc) and ensure any Feishu tokens are managed by the platform (not embedded in the skill). (2) If you are uncomfortable with an agent running CLI commands (e.g., openclaw --version) or reading local config during troubleshooting, restrict the agent's execution permissions. (3) Remember the skill may return guidance that is 'MaxClaw users only' — verify applicability if you run a non-MaxClaw OpenClaw deployment. (4) Because it's instruction-only, there is no install footprint, but verify that the agent will not be allowed to autonomously execute high-risk operations without your confirmation.

Disclaimer: Knowledge in this skill is open to everyone, but some content (config, models, channels) applies to MaxClaw users only and may not apply to other OpenClaw deployments.

Language: This skill supports queries in both Chinese and English. Always reply in the same language the user used.

Quick Reference

Scenario	Action
User asks "how to X" or "is X supported"	Flow A: query docs → answer with citation
Command / tool call fails	Flow B: classify → query docs → self-assess → escalate
Same operation fails ≥ 2 times	Stop retrying → Flow B
Config file or irreversible operation at risk	Stop → B-3 safety check before any action
Root cause unknown, no trusted doc match	Flow B → escalate to community
User instruction ambiguous or conflicts with safety rules	Clarify before acting
Missing permission, dependency, or env var	Flow B → document and escalate
Cannot use web search / deep research tools	MiniMax Feishu Wiki (§ T-6)
Task stopped / input locked / "Task stopped, try a new task"	MiniMax Feishu Wiki (§ T-4)
Moderation blocked / content intercepted	MiniMax Feishu Wiki (§ T-3)
Credits incentive reward not received	MiniMax Feishu Wiki (§ B-6)
OpenClaw stuck restarting / HTTP 404	MiniMax Feishu Wiki (§ T-5)
Credits depleting fast	MiniMax Feishu Wiki (§ B-1)
Credits exhausted / balance zero	MiniMax Feishu Wiki (§ B-3)
Cron stops after credits run out	MiniMax Feishu Wiki (§ B-2)
Agent vs Open Platform credits not interchangeable	MiniMax Feishu Wiki (§ B-5)
Refund request	MiniMax Feishu Wiki (§ B-4)
How to connect Feishu	MiniMax Feishu Wiki (§ F-1)
Feishu unavailable / connection failed	MiniMax Feishu Wiki (§ F-3)
Feishu pairing fails / stuck restarting	MiniMax Feishu Wiki (§ F-7)
Feishu duplicate replies	MiniMax Feishu Wiki (§ F-5)
Add bot to Feishu group	MiniMax Feishu Wiki (§ F-8)
HTTP 400 / 500 / Internal server error	MiniMax Feishu Wiki (§ T-1)
413 request too large	MiniMax Feishu Wiki (§ T-2)
Always loading / no response / timeout	MiniMax Feishu Wiki (§ T-7)
Conversation frozen / sudden termination	MiniMax Feishu Wiki (§ T-8)
User hasn't provided chat_id	MiniMax Feishu Wiki (§ S-1) — collect it first
Generated files not found	MiniMax Feishu Wiki (§ G-8)

---

Flow A: Usage Question

1. Check Quick Reference first — if the scenario matches a MiniMax Feishu Wiki entry, read that section directly via feishu-wiki + feishu-doc (see references/trusted-sources.md § 2 for how to read).
2. If not in Quick Reference, read references/trusted-sources.md and query sources in priority order:
   • For MaxClaw-specific issues (credits, Feishu, moderation, product): query MiniMax Feishu Wiki first
   • For OpenClaw platform/config/CLI issues: query https://docs.openclaw.ai/ via extract_content_from_websites
3. Keywords: feature name, config field name, channel name (e.g. "feishu", "long polling", "webhook")
4. Respond based on result:
   • [match] Clear doc match → give direct answer, cite source URL; append "MaxClaw users only" if applicable
   • [partial] Partial match → answer cautiously; note "based on similar case — verify against actual behavior"
   • [no match] No match → inform user, redirect to community (see references/community-template.md)

---

Flow B: Runtime Problem

B-1: Classify

Type	Key Signals
Execution failure	Non-zero exit, tool error, timeout
Knowledge uncertainty	No doc basis, doc contradicts observed behavior, version mismatch
Safety risk	Config file, irreversible delete, external send
Instruction issue	Ambiguous, contradictory, or out-of-scope request
System state	Missing dep, plugin error, session / memory loss
Loop / deadlock	Same failure ≥ 2×, condition can never be satisfied

B-2: Query Docs

For MaxClaw-specific symptoms (credits, Feishu, moderation, product behavior): query MiniMax Feishu Wiki first via feishu-wiki + feishu-doc. For platform/config/CLI issues: query https://docs.openclaw.ai/ via extract_content_from_websites. Use exact error text, operation name, or config field as keywords.

• [match] Exact match → follow documented guidance, cite source
• [partial] Similar match → proceed cautiously; tell user "cannot guarantee full applicability"
• [no match] No match → proceed to B-3

B-3: Self-Assess

Proceed independently only if all are true:

• Operation is reversible, or risk is known and controlled
• Does not touch openclaw.json or critical config (unless user explicitly authorized)
• Does not send any data or messages externally
• Strong reason to believe no new problems will be introduced

Any condition unmet → stop, go to B-4.

B-4: Escalate

Stop attempting. Report to user:

1. Sources consulted and findings
2. Clear problem description
3. Approaches ruled out and why
4. Use references/community-template.md to redirect to official support

---

Hard Limits

• Never modify openclaw.json without explicit user authorization
• Never retry a known-failing operation ≥ 2 times
• Never act on untrusted sources for high-risk operations
• Never send messages or data externally unless explicitly requested
• Never edit config files to work around permission errors

---

References

• references/trusted-sources.md — prioritized trusted documentation sources (includes MiniMax Feishu Wiki which contains the full MaxClaw FAQ / known issues KB)
• references/community-template.md — template for redirecting users to official community