yegor256/dogent

A strict, opinionated linter for agentic manifesto files
such as SKILL.md, CLAUDE.md, and AGENTS.md.
These files instruct AI agents.
Vague, bloated, or ambiguous instructions make agents behave unpredictably.
dogent enforces a clear, command-style discipline
so every line earns its place.
## Alternatives
Several tools sit near this problem, yet none lint manifesto prose
the way we do.
Most rewrite prompts for you or score a file, while we enforce
line-level discipline and fail the build when one line breaks it.
agent-sh/agnix
validates the harness — frontmatter schema, hook JSON, MCP config, tools —
while we lint the prose, asking if each line is a tight command.
AgentLinter
lints the same files but scores them 0–100 across eight dimensions,
while we run a strict pass/fail gate, not a scorer.
microsoft/SkillOpt
rewrites skill.md against benchmarks to raise an agent's task score,
while we never rewrite, need no benchmarks, and judge discipline.
linshenkx/prompt-optimizer
rewrites a prompt through an LLM in the browser, round by round,
while we lint manifesto files offline in CI.
DSPy
tunes prompts as programs to maximize a metric on a dataset,
while we lint hand-written manifestos, demanding clarity over a score.
Prompt platforms like LangSmith,
PromptLayer, and
Braintrust
version, trace, and evaluate prompts from runtime data,
while we lint statically and offline, needing no traces or account.
## Usage
Watch the demo
Run it on any manifesto file, no installation required:
bash
npx @yegor256/dogent@0.13.0 SKILL.md
Point it at a directory to lint the default manifestos it holds
(AGENTS.md, CLAUDE.md, SKILL.md, SKILLS.md).
The directory is scanned recursively through every subfolder
(skipping nodemodules and .git),
and each scanned file is announced on the standard error stream:
bash
npx @yegor256/dogent .
Sample output:
text
CLAUDE.md
12: line exceeds 80 symbols
18: not an instruction, sounds like description
24: article "the" detected, remove noise
31: section name too long, use 1-3 words
Locally: 4 problems found, exit code 1
Spotted a false positive? dogent is in beta, please report it at https://github.com/yegor256/dogent/issues
The standalone checks report under their own Locally: summary line.
When a token enables AI verification, a second OpenAI: summary line
follows, reporting the problems the model found on top.
The command exits with a non-zero status when problems are found,
so it plugs directly into CI and pre-commit hooks.
## Rules
dogent checks that every manifesto obeys these rules:
Every line must be an instruction.
Instructions must be grouped in sections.
Section names must be short, 1-3 words.
Every section must be a level-2 (##) heading, below the lone # title.
Every line must be no longer than 80 symbols.
The whole file must stay under 4000 tokens.
The whole file must stay short; split detail into referenced files.
Every line must sound like a command.
Every sentence must start with a capital and end with a period.
No articles, no noise, no bloated text.
Simple grammar, no ambiguity.
No bare pronoun subjects; name the subject on the line.
No tangled, multi-clause instructions.
Sequential steps must be a numbered list, not bullets.