Wiki Compiler

AdvisoryAudited by Static analysis on May 13, 2026.

Overview

No suspicious patterns detected.

Findings (0)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

The agent may be able to read and modify IMA knowledge-base content according to the permissions of the supplied API key.

Why it was flagged

The skill requires IMA API credentials to access the user's IMA knowledge-base and note APIs.

Skill content
IMA_OPENAPI_CLIENTID: ima OpenAPI 客户端ID; IMA_OPENAPI_APIKEY: ima OpenAPI API密钥
Recommendation

Use a scoped or revocable IMA API key, avoid pasting credentials into shared logs, and revoke the key if you no longer use the skill.

What this means

The agent may perform API operations against a selected knowledge base; mistakes in target IDs or scope could affect the wrong workspace content.

Why it was flagged

The instructions show direct API calls and local command snippets for listing, exporting, and updating IMA wiki/note content.

Skill content
curl -s -X POST "https://ima.qq.com/openapi/wiki/v1/get_knowledge_list" ... python3 -c
Recommendation

Confirm the target knowledge_base_id or note_id before allowing write/update actions, and review generated content before it is saved.

What this means

Private or inaccurate source material could become part of a persistent guide and influence future answers.

Why it was flagged

The skill intentionally turns source materials into persistent wiki content that may be reused as a knowledge source in later work.

Skill content
编译后的 Wiki 是"真理之源"... 知识在系统中持续累积和演化。
Recommendation

Only include sources you want preserved in the wiki, review the generated guide for sensitive or incorrect content, and keep clear source attribution.

NoteMedium Confidence
ASI10: Rogue Agents
What this means

A user might assume future updates are automatic or already approved.

Why it was flagged

The wording suggests recurring maintenance behavior, although the artifacts contain no scheduler, daemon, or autonomous background code.

Skill content
第五步:主动维护与迭代 — 定期检查和更新
Recommendation

Require explicit user approval before each maintenance run or knowledge-base update.