Skip to content

docs: add agentchecker CLI for rule alignment #316

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
moisesvalero wants to merge 2 commits into
base: main
Choose a base branch
from
Open

docs: add agentchecker CLI for rule alignment #316

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
4c11d77
docs: add agentchecker to agent instruction tools
moisesvalero Jun 19, 2026
f9b843a
docs: add agent instruction alignment rule for agentchecker
moisesvalero Jun 21, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
40 changes: 40 additions & 0 deletions rules/agent-instruction-alignment-cursorrules-prompt-file.mdc
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,40 @@
---
description: "Cursor rules for keeping AGENTS.md, CLAUDE.md, Cursor rules, and Copilot instructions aligned across AI coding tools."
globs: AGENTS.md,**/AGENTS.md,CLAUDE.md,**/.claude/CLAUDE.md,.github/copilot-instructions.md,.cursor/rules/**.mdc
alwaysApply: false
---

# Agent instruction alignment

When editing any agent instruction file in this repository, keep stack and workflow decisions consistent across all surfaces.

## Canonical sources

Treat `AGENTS.md` as the project-wide contract for humans and shared agents. Mirror its decisions in:

- `.cursor/rules/*.mdc` for Cursor
- `CLAUDE.md` or `.claude/CLAUDE.md` for Claude Code
- `.github/copilot-instructions.md` for GitHub Copilot

## Keep these fields aligned

- Package manager (pnpm, npm, yarn, bun)
- Linter and formatter (oxlint vs eslint, prettier vs biome)
- Test runner and type checker
- Runtime version constraints
- Shell and environment expectations
- Commit and comment conventions

## Editing workflow

1. Pick one source of truth (usually `AGENTS.md`).
2. When changing stack or tooling rules, update every agent file in the same PR.
3. Use identical wording for critical constraints (for example "never use npm").
4. Run `npx agentchecker` before opening a PR to detect drift.
5. Use `npx agentchecker --check-only` in CI to block contradictory merges.
Comment on lines +33 to +34

@coderabbitai coderabbitai Bot Jun 21, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

⚠️ Potential issue | 🟡 Minor

🧩 Analysis chain

🌐 Web query:

agentchecker CLI --check-only flag support

💡 Result:

There are multiple distinct command-line interface (CLI) tools named agentchecker or agentcheck, and none of the widely documented versions explicitly support a --check-only flag [1][2][3][4][5]. Because agentchecker/agentcheck refers to different projects, you should verify which tool you are using: 1. AgentCheck (by hvardhan878 / PyPI agentcheck): This tool focuses on tracing, replaying, and asserting agent executions [4][5]. Its CLI supports commands like trace, replay, diff, assert, and show, but it does not list --check-only [4][5]. 2. AgentCheck (by Pringled): This tool scans for agent access to resources (e.g., IAM, API keys) [2]. Its CLI uses flags such as --ci and --fail-on to control behavior, but does not include --check-only [2]. 3. AgentCheck (by Eshaank08): This tool performs audits of agent instruction files [3][6]. It supports flags like --path, --interactive, --no-ai, and --json, but does not list --check-only [3][6]. 4. agentchecker (associated with cursorrules/AGENTS.md): This is a utility for scanning instruction files for contradictions [1]. Documentation for this tool is minimal and does not specify a --check-only flag [1]. If you are looking to perform a "check-only" operation, you may need to use an alternative approach specific to the tool you have installed, such as using --no-ai for static analysis only [3][6] or using assertions for CI-friendly validation [4][5].

Citations:

Remove or correct the --check-only flag reference in line 34.

The agentchecker tool does not support a --check-only flag. Verify the actual CLI options available for the agentchecker version you're using and either remove this flag or replace it with the correct flag (e.g., --no-ai for static-only analysis, depending on your specific tool).

🤖 Prompt for AI Agents 
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@rules/agent-instruction-alignment-cursorrules-prompt-file.mdc` around lines
33 - 34, The `npx agentchecker --check-only` command in line 34 references a
flag that does not exist in the agentchecker tool. Verify the actual available
CLI options for the agentchecker version being used, then either remove the
`--check-only` flag entirely or replace it with the correct flag that provides
the intended behavior (such as `--no-ai` for static-only analysis if that
matches your requirements). Update the instruction to reflect the accurate
command syntax.


## Anti-patterns

- Adding tool-specific exceptions without documenting them in all files
- Duplicating long prose — link to `AGENTS.md` instead
- Leaving stale instructions after migrating tooling
Loading