HumanLayer is a software system that runs coding agents on a team’s machines and coordinates their tasks, sessions, plans, and reviews through desktop and web apps. Its Turborepo monorepo contains the user interfaces, host daemon, API, command-line tools, synchronization services, and agent plugins. The catalogue entries provide skills, plugins, instructions, and settings for its coding-agent workflows.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add humanlayer/humanlayer --skill describe-prgit clone --depth 1 https://github.com/humanlayer/humanlayerWrote 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/skills/humanlayer/humanlayer/describe-pr)<a href="https://agentmods.dev/skills/humanlayer/humanlayer/describe-pr"><img src="https://agentmods.dev/badge/skills/humanlayer/humanlayer/describe-pr/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/humanlayer/humanlayer/describe-pr"><img src="https://agentmods.dev/badge/skills/humanlayer/humanlayer/describe-pr.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00015 | $0.01030 |
| Opus 5.5 | $0.00006 | $0.00412 |
| Sonnet 5.5 | $0.00003 | $0.00206 |
| Haiku 4.5 | $0.00002 | $0.00103 |
Grade A, and why
describe-pr 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 today.
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 — 62 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Describe a Pull Request
Create or update the pull request for the current task with a concise description that helps a reviewer understand why the change exists and the shape of the implementation.
Workflow
-
Read the description template:
Read({SKILLBASE}/references/pr_description_template.md) -
Identify or create the pull request:
- If
ghis missing or signed out, stop and tell the user. Do not work around it. - Check the current branch for a PR with
gh pr view --json url,number,title,state,baseRefName,headRefName 2>/dev/null. - If no PR exists, inspect
git status --short --branchand the commits on the current branch. - Commit task-related changes when needed, push the branch with an upstream, and create a PR for it. Follow the repository's git safety protocol.
- Ask the user to select a PR only when the current branch has no relevant work and there is no safe current-branch PR to create.
- If
-
Gather only the context needed to explain the change:
- Read the ticket and any relevant task artifacts.
- Read the complete PR diff and enough surrounding code to understand behavior and ownership.
- Use
gh pr viewto collect PR metadata and changed files. - Read
{SKILLBASE}/references/show-me.mdfor the visual-outline conventions used in the PR body.
-
Write the PR description using the template:
- Keep Why the change to exactly one sentence.
- Keep Special things to note to 1-3 bullets. Prioritize reviewer warnings, migrations, compatibility constraints, deliberate omissions, or surprising decisions. Write
- None.when there are no special considerations. - Make Change outline a compact,
/show-me-inspired structural view rather than prose or a file-by-file changelog. - Include only the views that help explain this PR:
- SQL table and endpoint contract changes, plus pseudocode for business logic.
- A shallow file tree showing changed responsibilities.
- React component tree changes, including important hooks, state, and package boundaries.
- Call-tree, call-stack, control-flow, or data-flow changes.
- Prefer
diffblocks when showing changes to an existing shape. Show the complete target shape when most of it is new or diff notation would obscure ownership or order. - Keep each view focused on the calls, files, fields, components, and boundaries a reviewer needs. Omit categories that did not change.
- Do not add walkthrough artifacts, exhaustive implementation narratives, verification checklists, changelog text, or plan-deviation sections.
What ships with it
6 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- today First seen · 62 lines · 15 tokens per session scan A bb4f18b8b7ac
describe-pr is a skill published in the GitHub repository humanlayer/humanlayer (11,717 stars, last pushed today), licensed Apache-2.0. It adds 15 tokens to every session and 1,030 once invoked, about $0.0001 per session on Opus 5.5. 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 2026-10-23.
Other skills, from other repositories
otto-pr-review-loop
Run Polly and OCR on a PR until all findings settle.
resolve-drive-pr
Drive an open PR through Polly/OCR, validation, and maintainer handoff.
feature-publish
Commit the feature, write the PR body, publish in the selected mode, and drive the PR.
otto-review-remediation
Handle trusted reviewers' change requests on Otto PRs.
otto-drive-pr
PR mergeability, CI, and live preview.
lean-review
CocoLean diff-scoped over-engineering audit — scans uncommitted git diff and applies five classification tags (delete/stdlib/native/yagni/shrink) to identify unnecessary surface area before commit.