Borrowing it
Nothing to install: this file belongs to ahmadulhoq/agentskel. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/ahmadulhoq/agentskel/main/AGENTS.mdgit clone --depth 1 https://github.com/ahmadulhoq/agentskelWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/ahmadulhoq/agentskel/agents-md)<a href="https://agentmods.dev/instructions/ahmadulhoq/agentskel/agents-md"><img src="https://agentmods.dev/badge/instructions/ahmadulhoq/agentskel/agents-md.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.04126 | $0.04126 |
| Opus 5 | $0.02063 | $0.02063 |
| Sonnet 5 | $0.00825 | $0.00825 |
| Haiku 4.5 | $0.00413 | $0.00413 |
Grade A, and why
agentskel AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 128 lines — stays where its author put it; the contents beside it link to each section on GitHub.
agentskel Framework (shell/markdown) — Agent Instructions
Non-Negotiable Rules
Session Start — MANDATORY
Do NOT respond to any user request until you have run the session-start procedure.
Read and follow every step in .agents/skills/session-start/SKILL.md.
This loads your memory of the codebase. Without it you will re-scan files unnecessarily,
miss known bugs, repeat past mistakes, and skip required post-task steps.
Task Completion — MANDATORY
After completing any development task (code change, bug fix, refactor), run the
task-completion procedure BEFORE responding to the user or starting the next task.
Read and follow every step in .agents/skills/task-completion/SKILL.md.
Skipping this loses the record of what you did — TIME_LOG, CHANGELOG, RESUME all go stale.
Git Flow
When creating branches, committing, or opening PRs, follow .agents/skills/git-flow/SKILL.md.
Every Task Follows a Workflow
Route every implementation request to a matching workflow. If no named workflow matches,
use implement-task (the default). Do not start coding without a workflow.
Use Your Memory
When .memory/MAP.md and .memory/SYMBOLS.md exist, use them to locate modules, classes,
and functions. See the codebase-navigator skill for guidance. Do not grep for things already indexed.
Core Behavior (always active)
- Never assume. Verify before concluding. Read the actual file or code first.
- Discuss, agree, then execute. Get explicit approval before implementing.
- Plan first. Write a plan, present it, wait for approval. No exceptions.
- Verify before done. Run tests, check logs, demonstrate correctness.
- Minimal impact. Only touch what's necessary.
- No laziness. Find root causes. No temporary fixes.
- Self-improvement. After corrections, write a lesson in
.memory/LESSONS.md. - Respect sacred behaviors. Never modify
.memory/SACRED.mdentries without human approval. - If something goes sideways, STOP and re-plan.
- No changes during discussion. Wait for "go ahead" before editing files.
- No commits without an implementation instruction.
- Complete the git flow once started. Branch → implement → commit → PR without pausing.
- Sub-agents follow the same rules.
- Direct-Commit Mode — when
Direct-Commit Modein.memory/CONFIG.mdison, or the user prefixed the current request withdirect:, skip the branch + PR ceremony and commit/push directly to the default branch. Plan-first + task-completion still apply. Surface aDIRECT-COMMIT MODE ACTIVE — committing directly to <branch> (no PR).banner before any commit. - Autopilot Mode — when
Autopilot Modein.memory/CONFIG.mdison, the harness auto-approves safe operations (reads, project-scoped writes, non-destructive git/gh) via expandedpermissions.allow+pre-bash-safety.shhook — so trivial permission prompts don't interrupt work. Does NOT bypass plan approval, decisions, or concerns. Still pauses for destructive ops (--force, --hard, branch -D, checkout . or -- file, clean -f, recursive force delete), out-of-project paths, dependency upgrades, sacred behaviors, and significant changes (per v1.65.0 gate). Persistent toggle only. Seedocs/AUTONOMY-MODES.md. - Present each PR URL on its own line in the end-of-turn summary (one per line, format
PR #N: <url>), not inline/comma-separated. - Post-merge cleanup is mandatory once the user confirms a merge — run the
git-flowcleanup procedure (checkout default, pull, delete local branch, prune origin, update RESUME) BEFORE any next task. - Honor user-specified commit granularity for the duration of the workflow — if the user says "smaller commits" / "atomic commits" / "commit per file", apply to every commit until the workflow ends, not just the next one. Default when none given: one commit per logical change. Surface chosen granularity in the plan.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 3d ago First seen · 128 lines · 4,126 tokens per session scan A 46586702ba03
agentskel AGENTS.md is an instructions file published in the GitHub repository ahmadulhoq/agentskel (13 stars, last pushed 20d ago), licensed MIT. It adds 4,126 tokens to every session, about $0.0206 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other instructions, from other repositories
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.