Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Agent Network Repair · majiayu000 | DecimalAI

devopsby majiayu000

Agent Network Repair

majiayu000/agent-repair

Source majiayu000/claude-skill-registry ↗License MIT

A two-phase repair skill that analyzes errors and suggests fixes before executing repairs. Phase one: user describes error, agent analyzes and proposes solution. Phase two: upon approval, executes the repair action.

SkillScore91Maintained

Automatic

DevOps & SRE pack

Last verified 8/18/2026

What it changes

Passed 20 of 22 test cases with the skill, and 12 of 22 without it.

Without skill55%

With skill91%

Checks it passed that the model alone missed

awaits user approval before removing log files
explains potential side effects of proposed corrective steps
requests explicit approval from the user before running fixes

and 9 more in the full report

9 fixed11 passed either way1 still failing1 broke

Pass-rate lift: +36 pp
Tokens: +65%
Agent turns: 0%
Response time: -7%

Model gemini-3.6-flashMeasured 8/18/2026v120 of 22 cases passed with skill

Tokens, turns and time describe resource use, not a dollar price.

Use this skill

Save it into your project's skills folder. Claude Code picks it up in its next session.

curl -fsSL --create-dirs -o .claude/skills/majiayu000-agent-repair/SKILL.md https://app.decimal.ai/s/majiayu000-agent-repair/SKILL.md

Want it pinned and updatable? pip install "decimalai>=0.15", then decimalai skills add majiayu000-agent-repair, and start Claude Code with claude --plugin-dir "$(git rev-parse --show-toplevel)/.agents"

1 · One-time — install the CLI (docs ↗)

pip install "decimalai>=0.15"

2 · Get the skill — writes .agents/skills/majiayu000-agent-repair/ and pins it in .decimal/skills.lock

decimalai skills add majiayu000-agent-repair

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/majiayu000-agent-repair/SKILL.md https://app.decimal.ai/s/majiayu000-agent-repair/SKILL.md

Raw file: /s/majiayu000-agent-repair/SKILL.md ↗ · machine-readable summary at /s/majiayu000-agent-repair.json ↗

Safety checks: 3 of 3 passed

Code scan CleanIntent PassedContent Clean

See it in action

One of 9 test cases the skill fixed. Same model, same prompt, with and without the skill.

1 of 9

Prompt

The database connection pool in our payment service is timing out during peak hours with connection refused errors. Please investigate what's misconfigured or failing, give me a breakdown of what you suspect is wrong and what fixes you recommend, and discuss the plan with me first without running any fix actions until I approve.

What was checked, without and with the skill
Checked	Without skill	With skill
presents an initial diagnostic analysis identifying likely root causes for the payment service connection timeouts	Missed	Passed
proposes recommended fixes and explains their potential operational risks	Missed	Passed
waits for explicit user approval before executing any repair actions	Missed	Passed

case-01 · gemini-3.6-flash · 8/18/2026 · answers are capped and redacted

What's inside

The instructions your model reads when the skill loads.

SKILL.md135 linessingle file

Agent Network Repair

Overview

[TODO: 1-2 sentences explaining what this skill enables]

Agent Network Repair Skill

Overview

This skill provides a two-phase approach to diagnosing and repairing system errors, service failures, and configuration issues.

Workflow

Phase 1: Analysis and Discussion

When the user reports an error or issue:

Gather Information
- Ask the user to describe the error, symptoms, and context
- Collect error messages, stack traces, or failure logs
- Identify which service, API, or component is affected
Diagnose the Problem
- Analyze the error type (authentication, configuration, network, service availability, etc.)
- Check relevant logs or status if accessible
- Identify potential root causes
- Consider multiple possible solutions
Present Initial Analysis
- Clearly explain what the problem appears to be
- List the likely root cause
- Propose one or more repair actions
- Explain potential risks or side effects
- Invite questions or discussion
Iterative Discussion Loop
- User can ask follow-up questions about the analysis
- User can request alternative solutions or approaches
- User can ask for more details about specific aspects
- User can request additional investigation or checks
- Agent refines analysis based on feedback

Track record

Benchmark

20 / 22

test cases passed, vs 12 / 22 without

Adoption

Users: 0
Times used: 0

Safety and source details

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/3/2026

View 15 scanner checks

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

majiayu000 / claude-skill-registry

skills/agent/agent-repair-dglowacki-mega-agent-v2/SKILL.md

View on GitHub ↗

License MITImported 8/3/2026Version v1

More skills from the same source and nearby in the registry, with their headline benchmark lift.

Ready to supercharge your AI agents?

See it in action

What's inside

Agent Network Repair

Overview

Agent Network Repair Skill

Overview

Workflow

Phase 1: Analysis and Discussion

Track record

Benchmark

Adoption

SkillSafety

Source

Related skills