You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
- Drop "What's new" / "What's in v0.2.0" sections; Status line now points at Releases
- Drop "How the skills work" section; What-is-this + the skills themselves explain it
- Merge duplicated workflow-side auth list from Prerequisites into Authentication
- Drop recompile-reverts-tweak callout (lives in auth.md)
- Wrap Repository layout in <details> and condense the tree
- Collapse Local development to just the command; fold Releases footer into Credits
Net: 220 → 116 lines.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Copy file name to clipboardExpand all lines: README.md
+31-89Lines changed: 31 additions & 89 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -2,29 +2,17 @@
2
2
3
3
A Claude Code plugin for conversational discovery and installation of GitHub agentic workflows (gh-aw), with subscription-aware auth setup.
4
4
5
-
> **Status**: v0.2.1 — polish release on top of v0.2.0's validated [agent-team pipeline](https://github.com/verkyyi/agent-team-playground/pull/5). End-user journey fixes, ~4-min speedup on installed workflows, and a test framework underneath.
6
-
7
-
## What's new in v0.2.1
8
-
9
-
-**Critical end-user journey fixes** — `workflow`-scope preflight in both install skills (caught a real first-time-user gotcha), TTY warning for `claude setup-token` in headless environments, stale "every push to main" docs cleaned up.
10
-
-**Agent-team pipelines ~4 min faster** — `safe-outputs.threat-detection: false` on all four agent-team workflows skips gh-aw's per-agent threat classifier (appropriate for trusted-input pipelines where the user labels their own issue).
11
-
-**Reviewer posts a pipeline-summary comment** on the issue after approve, with links to all four run pages + the PR. One jump-off point for the human.
12
-
-**Implementer prompt tuning** — trust-the-plan directive + 5-tool-call budget heuristic. Observed wall-clock: 20m54s → 6m41s on a comparable task (−68%).
13
-
-**MIT LICENSE file** added (was already declared in plugin.json).
14
-
-**Credits section** explicitly attributes gh-aw (GitHub Next), agentics (same team), and superpowers (Jesse Vincent) in README.
15
-
16
-
## What's in v0.2.0 (still)
17
-
18
-
-**`agent-team` pattern** — four workflows (spec → plan → impl → review) that collaborate on a single issue via `workflow_dispatch` handoffs. See [catalog/agent-team/](catalog/agent-team/README.md).
19
-
-**`/install-agent-team` skill** — atomic install of all four roles + OAuth tweak + seven labels. See [skills/install-agent-team/SKILL.md](skills/install-agent-team/SKILL.md).
5
+
> **Status**: v0.2.1 — see [Releases](https://github.com/verkyyi/github-agent-runner/releases) for the changelog.
20
6
21
7
## What is this?
22
8
23
9
`github-agent-runner` is a Claude Code plugin that helps you add AI-powered automation to any GitHub repository. It does two things:
24
10
25
-
1.**Discover** — recommends 1–3 agentic workflows from a curated catalog that match your repo's shape (language, CI setup, activity level, etc.).
11
+
1.**Discover** — recommends 1–3 agentic workflows from the [`githubnext/agentics`](https://github.com/githubnext/agentics) catalog that match your repo's shape (language, CI setup, activity level, etc.).
26
12
2.**Install** — walks you through fetching, authenticating, and wiring up each workflow end-to-end, including the OAuth token tweak that makes your Claude subscription work inside GitHub Actions.
27
13
14
+
It also ships the **agent-team** pattern — four workflows (spec → plan → impl → review) installable in one pass via `/install-agent-team`. See [catalog/agent-team/](catalog/agent-team/README.md).
15
+
28
16
## Quick start
29
17
30
18
Open your repo in Claude Code (or any coding agent) and paste:
@@ -47,44 +35,12 @@ The agent will add the marketplace, install the plugin, and run `/discover-workf
47
35
Both skill names are unique, so the short form works out-of-the-box. If another installed plugin ever ships the same skill name, prefix with the plugin name: `/github-agent-runner:discover-workflows`.
48
36
</details>
49
37
50
-
> Want to hack on the plugin itself? See [Local development](#local-development) below for the `claude --plugin-dir .` workflow.
51
-
52
-
## How the skills work
53
-
54
-
### `/discover-workflows`
55
-
56
-
Inspects your repo's shape (language, test presence, CI configuration, recent activity) using only local `git` and filesystem tools. Then fetches the current list of workflows from the upstream [`githubnext/agentics`](https://github.com/githubnext/agentics/tree/main/workflows) catalog at runtime — no local catalog to drift — reads frontmatter from the most promising candidates, and recommends up to 3 that fit, each with a one-sentence reason specific to your repo and an estimated setup friction level. Hands off directly to `/install-workflow` once you pick one.
57
-
58
-
All agentics workflows require Claude auth (OAuth or API-key). See [skills/install-workflow/auth.md](skills/install-workflow/auth.md) for the decision tree.
59
-
60
-
### `/install-workflow`
61
-
62
-
Takes a workflow name (or prompts you to run `/discover-workflows` first) and:
63
-
64
-
1. Checks that `gh` CLI and `gh aw` extension are available, and that you have write access
65
-
2. Asks once whether you have a Claude subscription or prefer the API-key path
66
-
3. Guides you through setting the appropriate repo secret — never echoes or stores your token
67
-
4. Runs `gh aw add <workflow>` to compile the `.lock.yml`
68
-
5. For the OAuth path: applies the required two-pass post-compile tweak and verifies the grep counts
69
-
6. Runs `gh aw validate` and summarizes every file changed
70
-
71
38
## Prerequisites
72
39
73
-
To use the plugin:
74
-
75
40
-[Claude Code](https://claude.ai/code) CLI installed and authenticated
- A Claude Pro, Max ($100), or Max ($200) subscription **or** an [Anthropic API key](https://console.anthropic.com)
82
-
- The appropriate secret set on the repository:
83
-
- OAuth path: `CLAUDE_CODE_OAUTH_TOKEN`
84
-
- API-key path: `ANTHROPIC_API_KEY`
85
-
- GitHub Discussions enabled if you install `weekly-research` (uses the "ideas" category)
86
-
87
-
See [skills/install-workflow/auth.md](skills/install-workflow/auth.md) for the complete auth decision tree.
43
+
- A Claude Pro/Max subscription **or** an [Anthropic API key](https://console.anthropic.com) for the installed workflows themselves — see [Authentication](#authentication).
88
44
89
45
## Authentication
90
46
@@ -99,8 +55,6 @@ Two paths are supported:
99
55
100
56
Full details — including the two-pass tweak rationale, verification grep counts, failure modes, and the ToS boundary explanation — are in [skills/install-workflow/auth.md](skills/install-workflow/auth.md).
101
57
102
-
**Important**: `gh aw compile` reverts the OAuth tweak in `.lock.yml` files. Re-apply [Steps 3–4 from auth.md](skills/install-workflow/auth.md#step-3--apply-the-post-compile-tweak-to-every-lockyml) after any recompile event (`gh aw compile`, `gh aw upgrade`, `gh aw fix`, or editing the `.md` source).
103
-
104
58
## Running on this repo
105
59
106
60
This repo dogfoods [`daily-repo-status`](.github/workflows/daily-repo-status.md), [`update-docs`](.github/workflows/update-docs.md), and [`weekly-research`](.github/workflows/weekly-research.md) — a live example of what `/install-workflow` sets up. See the `.github/workflows/` directory for the compiled `.lock.yml` files.
@@ -111,64 +65,52 @@ Beyond the one-workflow-per-job templates above, this repo ships reference patte
111
65
112
66
-**[agent-team](catalog/agent-team/README.md)** — four roles (spec → plan → impl → review) coordinating through structured comment blocks and a small internal label state machine. Install all four in one pass with `/install-agent-team`; dispatch tasks by opening an issue and adding a single `agent-team` label. Use when you want visible handoffs, human override between steps, and an audit trail per task.
shared/reporting.md # pulled verbatim from agentics
159
102
```
160
103
161
-
Changes to `skills/*/SKILL.md` take effect on the next Claude Code session reload. Changes to `.lock.yml` files can be validated at any time with `gh aw validate` (safe — does not recompile).
162
-
163
-
## Releases
164
-
165
-
See the [Releases tab](https://github.com/verkyyi/github-agent-runner/releases) for tagged versions and changelogs. Auto-discovered by [claude-plugins.dev](https://claude-plugins.dev); also listed on [ClaudePluginHub](https://claudepluginhub.com).
104
+
`.lock.yml` files are marked `linguist-generated` and `merge=ours` in `.gitattributes` to prevent spurious merge conflicts.
105
+
</details>
166
106
167
107
## Credits
168
108
169
-
This plugin is built on two open-source projects from the [GitHub Next](https://githubnext.com) team:
109
+
Built on two open-source projects from the [GitHub Next](https://githubnext.com) team:
170
110
171
111
-**[github/gh-aw](https://github.com/github/gh-aw)** — the agentic workflow compiler. Every workflow this plugin installs is a gh-aw `.md` source compiled by `gh aw add` / `gh aw compile`. Maintained by [@pelikhan](https://github.com/pelikhan), [@dsyme](https://github.com/dsyme), and others.
172
-
-**[githubnext/agentics](https://github.com/githubnext/agentics)** — the curated workflow catalog. `/discover-workflows` surfaces entries from this repo; every dogfooded workflow in `.github/workflows/` traces back to a source `.md` at `githubnext/agentics/workflows/`. The `shared/reporting.md` component under `.github/workflows/shared/` is also pulled from agentics and included verbatim.
112
+
-**[githubnext/agentics](https://github.com/githubnext/agentics)** — the curated workflow catalog. `/discover-workflows` surfaces entries from this repo; every dogfooded workflow in `.github/workflows/` traces back to a source `.md` at `githubnext/agentics/workflows/`.
173
113
174
114
Pattern inspiration from **[superpowers](https://github.com/obra/superpowers)** by [@obra](https://github.com/obra) — our agent-team spec → plan → impl → review loop mirrors the superpowers skill loop, reimplemented inline for headless gh-aw execution.
115
+
116
+
Plugin listed on [claude-plugins.dev](https://claude-plugins.dev) and [ClaudePluginHub](https://claudepluginhub.com).
0 commit comments