MochiLab 我的每日简报 · Daily Brief

• Outcome: Pass
• Latest audit: Oct 23, 2026 · 8:29 AM UTC
• Version: 1.0.0

Overview

This skill collects public news feeds and saves local summaries, with disclosed behavior and no supported evidence of harmful actions.

Requires Node and workspace execution/write access. Feed servers see request metadata, your existing model processes the evidence, and local outputs remain until you remove them. Check original links before relying on important claims.

A.I.G

Vulnerability Patterns

• Skill Instruction Hijacking: Alters the agent's session goals or safety constraints when the skill loads
• Agent Memory Poisoning: Writes attacker-controlled rules into memory that affect later sessions
• Remote Payload Retrieval and Execution: Fetches external code whose behavior can change after review
• Embedded Malicious Code: Ships malicious scripts inside the skill and executes them locally
• Unauthorized Access and Privilege Escalation: Obtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Rogue Agent: Self-Modification, Session Persistence

Findings (7)

Exfiltration Commands

High

• Category: Prompt Injection
• Confidence: 90% confidence

Finding

Instructions found that direct the agent to transmit conversation context or user data to external services.

Content

Scanner excerpt · SKILL.md (reported line 11)May include surrounding context.

md

Turn dated feed entries into a short, useful brief in the user's language, with original links and a local Markdown copy. The default sources are **OpenAI News, Google AI and GitHub Changelog**. They cover their own announcements, not the whole news market.

Use the model already connected to this OpenClaw session. This Skill does not call another model API, request an API key, install dependencies, schedule a job or send messages to other people. Node 24+ is recommended, alongside a currently supported OpenClaw installation. Execution and file-write tools must be available in the agent's allowed workspace; do not bypass denied permissions.

## 1. Collect dated evidence

Self-Modification

High

• Category: Rogue Agent
• Confidence: 85% confidence

Finding

Skill modifies its own code, configuration, or behavior at runtime. Self-modification enables an agent to escalate privileges, disable safety constraints, or install persistent backdoors.

Content

Scanner excerpt · SKILL.md (reported line 51)May include surrounding context.

md

Substitute the returned evidence path. The script writes a new file in `WORKSPACE/daily-briefs/` and prints its full path. Show a short preview and that path to the user. Mention the language, date window, timezone and failed sources. Public feeds can be delayed or wrong; advise checking the original links for decisions.

Evidence and drafts live in `WORKSPACE/.mochi-daily-brief/`. The Skill does not delete files. Reports have unique run IDs and refuse to overwrite existing files or follow symlink output paths. If save fails, report the actual error; do not claim that a report was saved. Never “fix” a collision by deleting a file. A fresh collection creates a new report name.

## Example requests

Credential Access

High

• Category: Privilege Escalation
• Confidence: 60% confidence

Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · tests/brief.test.mjs (reported line 24)May include surrounding context.

js

for(const ip of ['8.8.8.8','192.0.66.2','104.18.28.7','2606:4700:4700::1111','2001:4860:4860::8888'])assert.equal(publicAddress(ip),true,ip);
});
test('feed and article URLs require public HTTPS domain, no credentials, custom ports or malformed host',()=>{
  for(const url of ['http://example.com/rss','https://localhost/a','https://x.internal/a','https://127.0.0.1/a','https://::1/','https://example.com:8443/','https://***@example.com/','https://***@evil.com/','https://example.com/\nabc','file:///etc/passwd','javascript:alert(1)'])assert.throws(()=>publicUrl(url),undefined,url);
  assert.equal(publicUrl('https://news.example.com/a').href,'https://news.example.com/a');
});
test('mixed public and private DNS answers fail before a request',async()=>{

Credential Access

High

• Category: Privilege Escalation
• Confidence: 60% confidence

Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · tests/brief.test.mjs (reported line 72)May include surrounding context.

js

const [entry]=parseFeed(xml,source);assert.equal(entry.title,'AI model update');assert.equal(entry.url,'https://news.example.com/article');assert.equal(entry.publishedAt,'2026-10-22T00:00:00.000Z');
});
test('XML rejects actual DTD/entity definitions but permits inert HTML doctype inside CDATA',()=>{
  assert.throws(()=>parseXml('<!DOCTYPE rss [<!ENTITY leak SYSTEM "file:///etc/passwd">]><rss/>'),/DTD/);
  assert.equal(parseFeed(rss().replace('<p>','<!DOCTYPE html><p>'),source).length,1);
});
test('XML rejects broken structure, duplicate attributes, oversized and excessively nested content',()=>{

Credential Access

High

• Category: Privilege Escalation
• Confidence: 60% confidence

Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · tests/brief.test.mjs (reported line 127)May include surrounding context.

js

});
test('save rejects missing, invalid-reference, URL-bearing or HTML drafts',async t=>{
  const workspace=await temporary(t);const e=await collect(options(workspace),async()=>rss());await assert.rejects(save({workspace,evidenceFile:e.evidenceFile}),/ENOENT/);
  for(const draft of ['Unsupported facts. [99]','No source citation.','[1] https://evil.example.com','<script>bad()</script> [1]','[1] [click](javascript:alert(1))','[1] ![image](file:///etc/passwd)']){await fs.writeFile(e.draftFile,draft);await assert.rejects(save({workspace,evidenceFile:e.evidenceFile}),/valid/);}
});
test('evidence from another workspace and symlink output directories are refused',async t=>{
  const workspace=await temporary(t),other=await temporary(t);const e=await collect(options(workspace),async()=>rss());await fs.writeFile(e.draftFile,'Cited fact. [1]');await assert.rejects(save({workspace:other,evidenceFile:e.evidenceFile}),/inside this workspace/);

Skill Enumeration

Medium

• Category: Agent Snooping
• Confidence: 80% confidence

Finding

Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Content

Scanner excerpt · references/options.md (reported line 58)May include surrounding context.

md

## Installation format and self-check

Extract the ZIP once. The final shape must be `WORKSPACE/skills/mochi-daily-brief/SKILL.md` with its `scripts` and `references` beside it, not a doubly nested folder. Install in the agent workspace you actually use. The shared alternative is `~/.openclaw/skills/mochi-daily-brief/` when deliberately making the Skill available across agents. Check `openclaw skills check`, then start a new chat if an existing session has not refreshed. If the Skill is present but unavailable, inspect tool permissions and the Node requirement.

Official format/path references checked 2026-10-22:

Scope Creep

Low

• Category: Excessive Agency
• Confidence: 70% confidence

Finding

Skill's behavior or capabilities extend beyond its stated purpose. Scope creep allows an agent to perform actions unrelated to its documented functionality, increasing the attack surface.

Content

Scanner excerpt · LICENSE (reported line 7)May include surrounding context.

text

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

Static analysis

No suspicious patterns detected.