agentflow

Name: agentflow
Author: sickn33

44kstars6.5kforks

How to install agentflow?

sickn33/agentic-awesome-skills/agentflow

$npx -y skills add sickn33/agentic-awesome-skills --skill agentflow

Installs into the current project.

Summary

Orchestrate autonomous AI development pipelines through your Kanban board (Asana, GitHub Projects, Linear). Manages multi-worker Claude Code dispatch, deterministic quality gates, adversarial review, per-task cost tracking, and crash-proof pipeline execution.

Files · 1

View on GitHub

SKILL.md

1	# AgentFlow
2
3	## Overview
4
5	AgentFlow turns your existing Kanban board into a fully autonomous AI development pipeline. Instead of building custom orchestration infrastructure, it treats your project management tool (Asana, GitHub Projects, Linear) as a distributed state machine — tasks move through stages, AI agents read and write state via comments, and humans intervene through the same UI they already use.
6
7	The result is complete pipeline observability from your phone, free crash recovery (state lives in your PM tool, not in memory), and human override at any point by dragging a card.
8
9	## When to Use This Skill
10
11	- Use when you need to orchestrate multiple Claude Code workers across a full development lifecycle (build, review, test, integrate)
12	- Use when you want deterministic quality gates (tsc/eslint/tests) before AI review on AI-generated code
13	- Use when you want full pipeline visibility from your Kanban board or phone
14	- Use when running a solo or team project that needs autonomous task dispatch with cost tracking
15	- Use when you need crash-proof orchestration that survives session restarts
16
17	## Core Concepts
18
19	### 7-Stage Kanban Pipeline
20
21	Tasks flow through: Backlog, Research, Build, Review, Test, Integrate, Done. Each stage has specific gates. The Kanban board IS the orchestration layer — no separate database, no message queue, no custom infrastructure.
22
23	### Stateless Orchestrator
24
25	A crontab-driven one-shot sweep runs every 15 minutes. No daemon, no session dependency. If it crashes, the next sweep picks up where it left off because all state lives in your PM tool.
26
27	### Deterministic Before Probabilistic
28
29	Hard gates (tsc + eslint + tests) run before any AI review, catching roughly 60% of issues at near-zero cost. AI review comes after, as a second layer.
30
31	### Adversarial Review
32
33	A different AI agent reviews code and must list 3 things wrong before deciding to pass. This prevents rubber-stamp approvals.
34
35	### Transitive Priority Dispatch
36
37	Tasks that unblock the most downstream work get built first, automatically computing the critical path.
38
39	## Skills / Commands
40
41	### `/spec-to-board`
42	Decomposes a SPEC.md into atomic tasks on your Kanban board with dependencies mapped.
43
44	### `/sdlc-orchestrate`
45	Dispatches tasks to workers based on transitive priority and conflict detection. Runs as a crontab sweep.
46
47	### `/sdlc-worker --slot <N>`
48	Runs a worker in a terminal slot that picks up tasks, builds code, and creates PRs. Run 3-4 workers in parallel.
49
50	### `/sdlc-health`
51	Real-time pipeline status dashboard showing current stage, assigned agent, retry count, and accumulated cost for every task.
52
53	### `/sdlc-stop`
54	Graceful shutdown: active workers finish their current task, unstarted tasks return to Backlog.
55
56	## Step-by-Step Guide
57
58	### 1. Write Your Spec
59
60	Create a `SPEC.md` for your project describing what you want to build.
61
62	### 2. Decompose Into Tasks
63
64	```
65	claude -p "/spec-to-board"
66	```
67
68	This reads your SPEC.md, decomposes it into atomic tasks, maps dependencies, and creates them on your Kanban board.
69
70	### 3. Start Workers
71
72	Open 3-4 terminal windows, each as a worker slot:
73
74	```bash
75	# Terminal 2 — Builder
76	claude -p "/sdlc-worker --slot T2"
77
78	# Terminal 3 — Builder
79	claude -p "/sdlc-worker --slot T3"
80
81	# Terminal 4 — Reviewer
82	claude -p "/sdlc-worker --slot T4"
83
84	# Terminal 5 — Tester
85	claude -p "/sdlc-worker --slot T5"
86	```
87
88	### 4. Start the Orchestrator
89
90	```bash
91	# Add to crontab (runs every 15 minutes)
92	crontab -e
93	# Add: /15 * * * ~/.claude/sdlc/agentflow-cron.sh >> /tmp/agentflow-orchestrate.log 2>&1
94	```
95
96	### 5. Monitor and Intervene
97
98	Open your Kanban board on your phone. Watch tasks flow through the pipeline. Drag any card to "Needs Human" to intervene. Run `/sdlc-health` for a terminal dashboard.
99
100	### 6. Stop the Pipeline
101
102	```
103	claude -p "/sdlc-stop"
104	```
105
106	## Quality Gates
107
108	Each stage enforces specific gates before promotion:
109
110	- Build to Review: `tsc` + `eslint` + `npm test` must all pass (deterministic)
111	- Review to Test: Adversarial reviewer must list 3 issues before passing
112	- Test to Integrate: 80% coverage threshold on new files
113	- Integrate to Done: Full test suite on main after merge; auto-reverts on failure
114
115	## Cost Tracking
116
117	Per-task cost tracking with stage ceilings (Sonnet defaults):
118
119	- Research: ~$0.10
120	- Build: ~$0.40
121	- Review: ~$0.10
122	- Test: ~$0.05
123	- Integrate: ~$0.03
124
125	Automatic guardrails: warning at $3/$8, hard stop at $10/$20 (Sonnet/Opus) with human escalation.
126
127	## Safety and Recovery
128
129	- Auto-revert: Integration failures trigger `git revert` (new commit, never force-push)
130	- Blocked tasks: After 2 failed attempts, tasks escalate to human revi

People who liked this also viewed

6 picks

Type

agentflow

44kstars6.5kforks

View on GitHub

How to install agentflow?

sickn33/agentic-awesome-skills/agentflow

$npx -y skills add sickn33/agentic-awesome-skills --skill agentflow

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# AgentFlow
2
3	## Overview
4
5	AgentFlow turns your existing Kanban board into a fully autonomous AI development pipeline. Instead of building custom orchestration infrastructure, it treats your project management tool (Asana, GitHub Projects, Linear) as a distributed state machine — tasks move through stages, AI agents read and write state via comments, and humans intervene through the same UI they already use.
6
7	The result is complete pipeline observability from your phone, free crash recovery (state lives in your PM tool, not in memory), and human override at any point by dragging a card.
8
9	## When to Use This Skill
10
11	- Use when you need to orchestrate multiple Claude Code workers across a full development lifecycle (build, review, test, integrate)
12	- Use when you want deterministic quality gates (tsc/eslint/tests) before AI review on AI-generated code
13	- Use when you want full pipeline visibility from your Kanban board or phone
14	- Use when running a solo or team project that needs autonomous task dispatch with cost tracking
15	- Use when you need crash-proof orchestration that survives session restarts
16
17	## Core Concepts
18
19	### 7-Stage Kanban Pipeline
20
21	Tasks flow through: Backlog, Research, Build, Review, Test, Integrate, Done. Each stage has specific gates. The Kanban board IS the orchestration layer — no separate database, no message queue, no custom infrastructure.
22
23	### Stateless Orchestrator
24
25	A crontab-driven one-shot sweep runs every 15 minutes. No daemon, no session dependency. If it crashes, the next sweep picks up where it left off because all state lives in your PM tool.
26
27	### Deterministic Before Probabilistic
28
29	Hard gates (tsc + eslint + tests) run before any AI review, catching roughly 60% of issues at near-zero cost. AI review comes after, as a second layer.
30
31	### Adversarial Review
32
33	A different AI agent reviews code and must list 3 things wrong before deciding to pass. This prevents rubber-stamp approvals.
34
35	### Transitive Priority Dispatch
36
37	Tasks that unblock the most downstream work get built first, automatically computing the critical path.
38
39	## Skills / Commands
40
41	### `/spec-to-board`
42	Decomposes a SPEC.md into atomic tasks on your Kanban board with dependencies mapped.
43
44	### `/sdlc-orchestrate`
45	Dispatches tasks to workers based on transitive priority and conflict detection. Runs as a crontab sweep.
46
47	### `/sdlc-worker --slot <N>`
48	Runs a worker in a terminal slot that picks up tasks, builds code, and creates PRs. Run 3-4 workers in parallel.
49
50	### `/sdlc-health`
51	Real-time pipeline status dashboard showing current stage, assigned agent, retry count, and accumulated cost for every task.
52
53	### `/sdlc-stop`
54	Graceful shutdown: active workers finish their current task, unstarted tasks return to Backlog.
55
56	## Step-by-Step Guide
57
58	### 1. Write Your Spec
59
60	Create a `SPEC.md` for your project describing what you want to build.
61
62	### 2. Decompose Into Tasks
63
64	```
65	claude -p "/spec-to-board"
66	```
67
68	This reads your SPEC.md, decomposes it into atomic tasks, maps dependencies, and creates them on your Kanban board.
69
70	### 3. Start Workers
71
72	Open 3-4 terminal windows, each as a worker slot:
73
74	```bash
75	# Terminal 2 — Builder
76	claude -p "/sdlc-worker --slot T2"
77
78	# Terminal 3 — Builder
79	claude -p "/sdlc-worker --slot T3"
80
81	# Terminal 4 — Reviewer
82	claude -p "/sdlc-worker --slot T4"
83
84	# Terminal 5 — Tester
85	claude -p "/sdlc-worker --slot T5"
86	```
87
88	### 4. Start the Orchestrator
89
90	```bash
91	# Add to crontab (runs every 15 minutes)
92	crontab -e
93	# Add: /15 * * * ~/.claude/sdlc/agentflow-cron.sh >> /tmp/agentflow-orchestrate.log 2>&1
94	```
95
96	### 5. Monitor and Intervene
97
98	Open your Kanban board on your phone. Watch tasks flow through the pipeline. Drag any card to "Needs Human" to intervene. Run `/sdlc-health` for a terminal dashboard.
99
100	### 6. Stop the Pipeline
101
102	```
103	claude -p "/sdlc-stop"
104	```
105
106	## Quality Gates
107
108	Each stage enforces specific gates before promotion:
109
110	- Build to Review: `tsc` + `eslint` + `npm test` must all pass (deterministic)
111	- Review to Test: Adversarial reviewer must list 3 issues before passing
112	- Test to Integrate: 80% coverage threshold on new files
113	- Integrate to Done: Full test suite on main after merge; auto-reverts on failure
114
115	## Cost Tracking
116
117	Per-task cost tracking with stage ceilings (Sonnet defaults):
118
119	- Research: ~$0.10
120	- Build: ~$0.40
121	- Review: ~$0.10
122	- Test: ~$0.05
123	- Integrate: ~$0.03
124
125	Automatic guardrails: warning at $3/$8, hard stop at $10/$20 (Sonnet/Opus) with human escalation.
126
127	## Safety and Recovery
128
129	- Auto-revert: Integration failures trigger `git revert` (new commit, never force-push)
130	- Blocked tasks: After 2 failed attempts, tasks escalate to human revi

People who liked this also viewed

6 picks

Type