Coding Agents

MegaLinter is designed to work hand in hand with your coding agent:

Get started

Install the MegaLinter agent skills at the root of your repository:

With Claude Code (the skills are copied directly into .claude/skills/):

npx skills add oxsecurity/megalinter/skills -s '*' -a claude-code -y

With another coding agent, replace claude-code with your agent's identifier (cursor, github-copilot, codex, antigravity, opencode... full list in the skills CLI documentation) — or let the CLI detect your installed agents and install the skills for all of them at once:

npx skills add oxsecurity/megalinter/skills -s '*' -y --copy

Then just talk to your agent:

"Setup MegaLinter on this repo"
"Run MegaLinter and fix the errors"
"Why is the MegaLinter CI job failing? Fix it"

What the skills do

Skill	Role
megalinter	Entry point: detects the repository state and orchestrates the other skills in a check → fix → re-check loop (3 iterations max)
megalinter-setup	Installs or upgrades MegaLinter using `npx mega-linter-runner --install` (non-interactive), then refines `.mega-linter.yml`
megalinter-check	Collects errors: watches a MegaLinter CI job (GitHub Actions, GitLab CI, Azure Pipelines, Bitbucket Pipelines) or runs MegaLinter locally with docker/podman
megalinter-fix	Fixes errors guided by a per-linter fix guide (one for each of the 124 linters, loaded only when its linter has errors), asks you before disabling rules or pushing

Sub-agents orchestration

On agents supporting sub-agents (Claude Code, OpenCode, GitHub Copilot custom agents...), megalinter-setup installs three sub-agents that keep large CI logs and linter outputs out of your main agent's context, and run on low-cost models when possible:

flowchart TD
    U(["👤 You: <i>run megalinter and fix the errors</i>"]) --> O["🎯 megalinter<br/>(orchestrator skill)"]
    O -->|"no config found"| S["🛠️ megalinter-setup"]
    S -->|"npx mega-linter-runner --install"| CFG[/".mega-linter.yml<br/>+ CI workflow"/]
    O --> C["🔍 megalinter-check"]
    C -->|"watch mode"| W[["👀 megalinter-watcher<br/>sub-agent, low-cost model"]]
    W --> CI[/"CI job logs<br/>GitHub / GitLab / Azure / Bitbucket"/]
    C -->|"local mode"| RUN[["🐳 megalinter-runner<br/>sub-agent, low-cost model"]]
    RUN --> IMG[/"MegaLinter Docker image<br/>docker or podman"/]
    W --> ERR[/"📋 Compact error list<br/>grouped by linter"/]
    RUN --> ERR
    ERR --> F["🔧 megalinter-fix"]
    F -->|"one sub-agent per failing linter,<br/>in parallel"| FX1[["🤖 megalinter-fixer<br/>+ python_ruff fix guide"]]
    F --> FX2[["🤖 megalinter-fixer<br/>+ markdown_markdownlint fix guide"]]
    FX1 --> RES[/"✏️ Fixed files<br/>+ proposed rule disables"/]
    FX2 --> RES
    RES -->|"you confirm disables"| RC["♻️ Targeted re-check<br/>parallel standalone linter images"]
    RC -->|"errors remain (max 3 loops)"| F
    RC -->|"clean"| DONE(["✅ Commit on a branch<br/>(never on main)"])

    classDef user fill:#e8f5e9,stroke:#2e7d32,stroke-width:2px,color:#1b5e20
    classDef skill fill:#ede7f6,stroke:#5e35b1,stroke-width:2px,color:#311b92
    classDef subagent fill:#e3f2fd,stroke:#1976d2,stroke-width:2px,color:#0d47a1
    classDef artifact fill:#fff8e1,stroke:#f9a825,stroke-width:1px,color:#5d4037
    classDef recheck fill:#e0f2f1,stroke:#00897b,stroke-width:2px,color:#004d40

    class U,DONE user
    class O,S,C,F skill
    class W,RUN,FX1,FX2 subagent
    class CFG,CI,IMG,ERR,RES artifact
    class RC recheck

The fix guides combine information generated from the linter descriptors (auto-fix support, rules documentation URLs, MegaLinter tuning variables) with curated fix, inline-disable and ignore instructions grounded in each linter's official documentation.

Safety rules

Safe fixes are applied automatically; ambiguous ones are asked to you
Disabling a linter or a rule always requires your confirmation
Commits are never pushed to the default branch