honest-agent — Claude Code skill: what it does, how to install
honest-agent finds your AI agent config files and appends directives that cut out sycophancy in favor of direct feedback.
no license ★ 301 updated 2026-10-23
npx skills add hoodini/ai-agents-skills --skill honest-agent -a claude-codeWhat honest-agent does
This setup skill searches the workspace for configuration files used by tools like Claude Code, GitHub Copilot, Cursor, Windsurf, Cline, Aider, and Continue.dev. Once it identifies existing setups, it asks whether to apply changes locally to the current project, globally across the user profile, or both.
Integrity of your existing setup is preserved through an append-only rule. The skill inspects the target file first, preserves existing rules, and appends the new style guidelines at the bottom. Output adjusts automatically to match the target environment, using markdown sections or plain-text blocks such as .cursorrules.
Restarting your IDE or terminal session right after execution is necessary for the updated guidelines to take effect.
What it helps with
- stop AI coding tools from flattering flawed technical decisions
- append constructive pushback instructions across multiple agent configs
- enforce direct technical criticism instead of conversational validation
- configure project-level and global honesty rules in a single run
Example prompts for the agent
- Set up honest agent rules for this repository
- Scan my workspace and append objective feedback directives to all detected agents
- Configure brutal honesty mode across project and global instructions
How to install honest-agent
1Skills CLI (Vercel)
npx skills add hoodini/ai-agents-skills --skill honest-agent -a claude-code2Manually: put the files in place
~/.claude/skills/honest-agent/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/honest-agent/
1Skills CLI (Vercel)
npx skills add hoodini/ai-agents-skills --skill honest-agent -a codex2Manually: put the files in place
.agents/skills/honest-agent/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add hoodini/ai-agents-skills --skill honest-agent -a cursor2Manually: put the files in place
.cursor/skills/honest-agent/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install hoodini/ai-agents-skills skills/honest-agent2Manually: put the files in place
.github/skills/honest-agent/SKILL.md1Skills CLI (Vercel)
npx skills add hoodini/ai-agents-skills --skill honest-agent -a gemini-cli2Platform CLI
gemini skills install https://github.com/hoodini/ai-agents-skills --path skills/honest-agentManually: put the files in place
.devin/skills/honest-agent/SKILL.mdThe legacy .windsurf/skills/ path still works
Claude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about honest-agent
How to install honest-agent in Claude Code?
Run npx skills add hoodini/ai-agents-skills --skill honest-agent -a claude-code in your terminal, or place the skill definition file at ~/.claude/skills/honest-agent/SKILL.md.
What does honest-agent do?
It scans your directory for instruction files like .claude/CLAUDE.md or .cursorrules, then appends guidelines telling the model to challenge assumptions and offer unfiltered criticism.
Does honest-agent overwrite custom instructions already present?
No, it follows an append-only pattern. It reads existing content first and places the communication rules at the end of the file.