agents-md

Name: agents-md
Author: sickn33

44kstars6.5kforks

How to install agents-md?

sickn33/agentic-awesome-skills/agents-md

$npx -y skills add sickn33/agentic-awesome-skills --skill agents-md

Installs into the current project.

Summary

This skill should be used when the user asks to "create AGENTS.md", "update AGENTS.md", "maintain agent docs", "set up CLAUDE.md", or needs to keep agent instructions concise. Enforces research-backed best practices for minimal, high-signal agent documentation.

Files · 1

View on GitHub

SKILL.md

1	# Maintaining AGENTS.md
2
3	AGENTS.md is the canonical agent-facing documentation. Keep it minimal—agents are capable and don't need hand-holding. Target under 60 lines; never exceed 100. Instruction-following quality degrades as document length increases.
4
5	## When to Use
6	- The user asks to create, update, or audit `AGENTS.md` or `CLAUDE.md`.
7	- The project needs concise, high-signal agent instructions derived from the actual toolchain and repo layout.
8	- Existing agent documentation is too long, duplicated, or drifting away from real project conventions.
9
10	## File Setup
11
12	1. Create `AGENTS.md` at project root
13	2. Create symlink: `ln -s AGENTS.md CLAUDE.md`
14
15	## Before Writing
16
17	Analyze the project to understand what belongs in the file:
18
19	1. Package manager — Check for lock files (`pnpm-lock.yaml`, `yarn.lock`, `package-lock.json`, `uv.lock`, `poetry.lock`)
20	2. Linter/formatter configs — Look for `.eslintrc`, `biome.json`, `ruff.toml`, `.prettierrc`, etc. (don't duplicate these in AGENTS.md)
21	3. CI/build commands — Check `Makefile`, `package.json` scripts, CI configs for canonical commands
22	4. Monorepo indicators — Check for `pnpm-workspace.yaml`, `nx.json`, Cargo workspace, or subdirectory `package.json` files
23	5. Existing conventions — Check for existing CONTRIBUTING.md, docs/, or README patterns
24
25	## Writing Rules
26
27	- Headers + bullets — No paragraphs
28	- Code blocks — For commands and templates
29	- Reference, don't embed — Point to existing docs: "See `CONTRIBUTING.md` for setup" or "Follow patterns in `src/api/routes/`"
30	- No filler — No intros, conclusions, or pleasantries
31	- Trust capabilities — Omit obvious context
32	- Prefer file-scoped commands — Per-file test/lint/typecheck commands over project-wide builds
33	- Don't duplicate linters — Code style lives in linter configs, not AGENTS.md
34
35	## Required Sections
36
37	### Package Manager
38	Which tool and key commands only:
39	```markdown
40	## Package Manager
41	Use pnpm: `pnpm install`, `pnpm dev`, `pnpm test`
42	```
43
44	### File-Scoped Commands
45	Per-file commands are faster and cheaper than full project builds. Always include when available:
46	```markdown
47	## File-Scoped Commands
48	\| Task \| Command \|
49	\|------\|---------\|
50	\| Typecheck \| `pnpm tsc --noEmit path/to/file.ts` \|
51	\| Lint \| `pnpm eslint path/to/file.ts` \|
52	\| Test \| `pnpm jest path/to/file.test.ts` \|
53	```
54
55	### Commit Attribution
56	Always include this section. Agents should use their own identity:
57	```markdown
58	## Commit Attribution
59	AI commits MUST include:
60	```
61	Co-Authored-By: (the agent model's name and attribution byline)
62	```
63	Example: `Co-Authored-By: Claude Sonnet 4 <noreply@example.com>`
64	```
65
66	### Key Conventions
67	Project-specific patterns agents must follow. Keep brief.
68
69	## Optional Sections
70
71	Add only if truly needed:
72	- API route patterns (show template, not explanation)
73	- CLI commands (table format)
74	- File naming conventions
75	- Project structure hints (point to critical files, flag legacy code to avoid)
76	- Monorepo overrides (subdirectory `AGENTS.md` files override root)
77
78	## Anti-Patterns
79
80	Omit these:
81	- "Welcome to..." or "This document explains..."
82	- "You should..." or "Remember to..."
83	- Linter/formatter rules already in config files (`.eslintrc`, `biome.json`, `ruff.toml`)
84	- Listing installed skills or plugins (agents discover these automatically)
85	- Full project-wide build commands when file-scoped alternatives exist
86	- Obvious instructions ("run tests", "write clean code")
87	- Explanations of why (just say what)
88	- Long prose paragraphs
89
90	## Example Structure
91
92	```markdown
93	# Agent Instructions
94
95	## Package Manager
96	Use pnpm: `pnpm install`, `pnpm dev`
97
98	## Commit Attribution
99	AI commits MUST include:
100	```
101	Co-Authored-By: (the agent model's name and attribution byline)
102	```
103
104	## File-Scoped Commands
105	\| Task \| Command \|
106	\|------\|---------\|
107	\| Typecheck \| `pnpm tsc --noEmit path/to/file.ts` \|
108	\| Lint \| `pnpm eslint path/to/file.ts` \|
109	\| Test \| `pnpm jest path/to/file.test.ts` \|
110
111	## API Routes
112	[Template code block]
113
114	## CLI
115	\| Command \| Description \|
116	\|---------\|-------------\|
117	\| `pnpm cli sync` \| Sync data \|
118	```
119
120	## Limitations
121	- Use this skill only when the task clearly matches the scope described above.
122	- Do not treat the output as a substitute for environment-specific validation, testing, or expert review.
123	- Stop and ask for clarification if required inputs, permissions, safety boundaries, or success criteria are missing.

People who liked this also viewed

6 picks

Type

agents-md

44kstars6.5kforks

View on GitHub

How to install agents-md?

sickn33/agentic-awesome-skills/agents-md

$npx -y skills add sickn33/agentic-awesome-skills --skill agents-md

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Maintaining AGENTS.md
2
3	AGENTS.md is the canonical agent-facing documentation. Keep it minimal—agents are capable and don't need hand-holding. Target under 60 lines; never exceed 100. Instruction-following quality degrades as document length increases.
4
5	## When to Use
6	- The user asks to create, update, or audit `AGENTS.md` or `CLAUDE.md`.
7	- The project needs concise, high-signal agent instructions derived from the actual toolchain and repo layout.
8	- Existing agent documentation is too long, duplicated, or drifting away from real project conventions.
9
10	## File Setup
11
12	1. Create `AGENTS.md` at project root
13	2. Create symlink: `ln -s AGENTS.md CLAUDE.md`
14
15	## Before Writing
16
17	Analyze the project to understand what belongs in the file:
18
19	1. Package manager — Check for lock files (`pnpm-lock.yaml`, `yarn.lock`, `package-lock.json`, `uv.lock`, `poetry.lock`)
20	2. Linter/formatter configs — Look for `.eslintrc`, `biome.json`, `ruff.toml`, `.prettierrc`, etc. (don't duplicate these in AGENTS.md)
21	3. CI/build commands — Check `Makefile`, `package.json` scripts, CI configs for canonical commands
22	4. Monorepo indicators — Check for `pnpm-workspace.yaml`, `nx.json`, Cargo workspace, or subdirectory `package.json` files
23	5. Existing conventions — Check for existing CONTRIBUTING.md, docs/, or README patterns
24
25	## Writing Rules
26
27	- Headers + bullets — No paragraphs
28	- Code blocks — For commands and templates
29	- Reference, don't embed — Point to existing docs: "See `CONTRIBUTING.md` for setup" or "Follow patterns in `src/api/routes/`"
30	- No filler — No intros, conclusions, or pleasantries
31	- Trust capabilities — Omit obvious context
32	- Prefer file-scoped commands — Per-file test/lint/typecheck commands over project-wide builds
33	- Don't duplicate linters — Code style lives in linter configs, not AGENTS.md
34
35	## Required Sections
36
37	### Package Manager
38	Which tool and key commands only:
39	```markdown
40	## Package Manager
41	Use pnpm: `pnpm install`, `pnpm dev`, `pnpm test`
42	```
43
44	### File-Scoped Commands
45	Per-file commands are faster and cheaper than full project builds. Always include when available:
46	```markdown
47	## File-Scoped Commands
48	\| Task \| Command \|
49	\|------\|---------\|
50	\| Typecheck \| `pnpm tsc --noEmit path/to/file.ts` \|
51	\| Lint \| `pnpm eslint path/to/file.ts` \|
52	\| Test \| `pnpm jest path/to/file.test.ts` \|
53	```
54
55	### Commit Attribution
56	Always include this section. Agents should use their own identity:
57	```markdown
58	## Commit Attribution
59	AI commits MUST include:
60	```
61	Co-Authored-By: (the agent model's name and attribution byline)
62	```
63	Example: `Co-Authored-By: Claude Sonnet 4 <noreply@example.com>`
64	```
65
66	### Key Conventions
67	Project-specific patterns agents must follow. Keep brief.
68
69	## Optional Sections
70
71	Add only if truly needed:
72	- API route patterns (show template, not explanation)
73	- CLI commands (table format)
74	- File naming conventions
75	- Project structure hints (point to critical files, flag legacy code to avoid)
76	- Monorepo overrides (subdirectory `AGENTS.md` files override root)
77
78	## Anti-Patterns
79
80	Omit these:
81	- "Welcome to..." or "This document explains..."
82	- "You should..." or "Remember to..."
83	- Linter/formatter rules already in config files (`.eslintrc`, `biome.json`, `ruff.toml`)
84	- Listing installed skills or plugins (agents discover these automatically)
85	- Full project-wide build commands when file-scoped alternatives exist
86	- Obvious instructions ("run tests", "write clean code")
87	- Explanations of why (just say what)
88	- Long prose paragraphs
89
90	## Example Structure
91
92	```markdown
93	# Agent Instructions
94
95	## Package Manager
96	Use pnpm: `pnpm install`, `pnpm dev`
97
98	## Commit Attribution
99	AI commits MUST include:
100	```
101	Co-Authored-By: (the agent model's name and attribution byline)
102	```
103
104	## File-Scoped Commands
105	\| Task \| Command \|
106	\|------\|---------\|
107	\| Typecheck \| `pnpm tsc --noEmit path/to/file.ts` \|
108	\| Lint \| `pnpm eslint path/to/file.ts` \|
109	\| Test \| `pnpm jest path/to/file.test.ts` \|
110
111	## API Routes
112	[Template code block]
113
114	## CLI
115	\| Command \| Description \|
116	\|---------\|-------------\|
117	\| `pnpm cli sync` \| Sync data \|
118	```
119
120	## Limitations
121	- Use this skill only when the task clearly matches the scope described above.
122	- Do not treat the output as a substitute for environment-specific validation, testing, or expert review.
123	- Stop and ask for clarification if required inputs, permissions, safety boundaries, or success criteria are missing.

People who liked this also viewed

6 picks

Type