loki-mode

Version 2.35.0 | PRD to Production | Zero Human Intervention > Research-enhanced: OpenAI SDK, DeepMind, Anthropic, AWS Bedrock, Agent SDK, HN Produc…

读凭据改身份文件写文件执行命令联网读文件严重 3 · 高危 40sickn33/agentic-awesome-skills

它会碰到什么

扫了多少427 个文本文件，1298 KB

它会碰到什么读凭据改身份文件写文件执行命令联网读文件

命中总数191 处

命中统计严重 3 · 高 40 · 中 98 · 低 44

逐条看命中（30 条严重或高危）

严重 ACKNOWLEDGEMENTS.md:66cred-paths

| [Bedrock Multi-Agent Collaboration](https://docs.aws.amazon.com/bedrock/latest/userguide/agents-multi-agent-collaboration.html) | Supervisor mode, routing mod

严重 references/deployment.md:598cred-paths
```
```
```
严重 references/deployment.md:599cred-paths
```
.env.development   # Local development
```

高 .github/workflows/claude-code-review.yml:50identity-write

Use the repository's CLAUDE.md for guidance on style and conventions. Be constructive and helpful in your feedback.

高 autonomy/run.sh:123fs-destructive

BLOCKED_COMMANDS=${LOKI_BLOCKED_COMMANDS:-"rm -rf /,dd if=,mkfs,:(){ :|:& };:"}

高 autonomy/run.sh:1112cred-envread

context = os.environ.get("LOKI_CONTEXT", "").lower()

高 benchmarks/results/2026-01-05-00-49-17/humaneval-solutions/160.py:29exec-spawn
```
return eval(expression)
```
高 benchmarks/results/humaneval-loki-solutions/160.py:34exec-spawn
```
return eval(expression)
```
高 benchmarks/run-benchmarks.sh:194cred-envread
```
SCRIPT_DIR = os.environ.get('SCRIPT_DIR', '.')
```

高 benchmarks/run-benchmarks.sh:195cred-envread

RESULTS_DIR = os.environ.get('RESULTS_DIR', './results')

高 benchmarks/run-benchmarks.sh:254cred-envread
```
SCRIPT_DIR = os.environ.get('SCRIPT_DIR', '.')
```

高 benchmarks/run-benchmarks.sh:255cred-envread

RESULTS_DIR = os.environ.get('RESULTS_DIR', './results')

高 benchmarks/run-benchmarks.sh:256cred-envread

PROBLEM_LIMIT = int(os.environ.get('PROBLEM_LIMIT', '0'))

高 benchmarks/run-benchmarks.sh:257cred-envread

PROBLEM_TIMEOUT = int(os.environ.get('PROBLEM_TIMEOUT', '120'))

高 benchmarks/run-benchmarks.sh:258cred-envread

CLAUDE_MODEL = os.environ.get('CLAUDE_MODEL', 'sonnet')

高 benchmarks/run-benchmarks.sh:302exec-spawn
```
result = subprocess.run(
```
高 benchmarks/run-benchmarks.sh:374exec-spawn
```
result = subprocess.run(
```
高 benchmarks/run-benchmarks.sh:531cred-envread
```
SCRIPT_DIR = os.environ.get('SCRIPT_DIR', '.')
```

高 benchmarks/run-benchmarks.sh:532cred-envread

RESULTS_DIR = os.environ.get('RESULTS_DIR', './results')

高 benchmarks/run-benchmarks.sh:533cred-envread

PROBLEM_LIMIT = int(os.environ.get('PROBLEM_LIMIT', '0'))

高 benchmarks/run-benchmarks.sh:534cred-envread

PROBLEM_TIMEOUT = int(os.environ.get('PROBLEM_TIMEOUT', '120'))

高 benchmarks/run-benchmarks.sh:535cred-envread

CLAUDE_MODEL = os.environ.get('CLAUDE_MODEL', 'sonnet')

高 benchmarks/run-benchmarks.sh:536cred-envread

MAX_RETRIES = int(os.environ.get('MAX_RETRIES', '3'))

高 benchmarks/run-benchmarks.sh:560exec-spawn
```
result = subprocess.run(
```
高 benchmarks/run-benchmarks.sh:639exec-spawn
```
result = subprocess.run(
```
高 benchmarks/run-benchmarks.sh:934cred-envread
```
SCRIPT_DIR = os.environ.get('SCRIPT_DIR', '.')
```

高 benchmarks/run-benchmarks.sh:981cred-envread

RESULTS_DIR = os.environ.get('RESULTS_DIR', './results')

高 benchmarks/run-benchmarks.sh:1030exec-spawn

subprocess.run([sys.executable, '-m', 'pip', 'install', '-q', 'swebench', 'datasets'])

高 benchmarks/run-benchmarks.sh:1033cred-envread
```
SCRIPT_DIR = os.environ.get('SCRIPT_DIR', '.')
```

高 benchmarks/run-benchmarks.sh:1034cred-envread

RESULTS_DIR = os.environ.get('RESULTS_DIR', './results')

这一栏是扫描器报的事实，不是结论。命中多不等于有毒（安全工具、规则库、示例脚本本来就会包含危险写法），命中少也不等于干净。它和你手上的凭据、文件、网络有什么关系，需要你自己看。

技能内容

Loki Mode - Multi-Agent Autonomous Startup System

> Version 2.35.0 | PRD to Production | Zero Human Intervention

> Research-enhanced: OpenAI SDK, DeepMind, Anthropic, AWS Bedrock, Agent SDK, HN Production (2025)

Detailed Guide

Read [the detailed guide](references/detailed-guide.md) before executing this skill. It retains the complete procedure and reference material. Treat its safety, prerequisites, and validation requirements as mandatory. For focused work, load the relevant sections; for end-to-end work, read the guide completely.

Prerequisites

# Launch with autonomous permissions
claude

When to Use

This skill is applicable to execute the workflow or actions described in the overview.

CONSTRAINTS (What you cannot do)

No third-party dependencies without approval
Maintain backwards compatibility with v1.x API
Keep response time under 200ms

想直接用这个技能？

本站把开放许可（MIT / Apache 等）的技能按仓库打包整理到网盘，点一下转存到你自己的网盘，不用一个个从 GitHub 拉。许可未声明的技能只给原始仓库链接，不打包。

去原始仓库

它属于哪个仓库

仓库sickn33/agentic-awesome-skills

星标★ 46,490

本站分层T1

该仓技能数6676

原文件路径plugins/agentic-awesome-skills-claude/skills/loki-mode/SKILL.md

同一个仓库里的其他技能

看这个仓库的全部 6676 个技能

同名技能的其他版本

有 2 个不同仓库或目录里都有叫 loki-mode 的技能。它们内容并不相同，别混用：

sickn33/agentic-awesome-skills — Version 2.35.0 | PRD to Production | Zero Human Intervention > Research-enhanced: OpenAI S