kyro-ai 3.2.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/.claude-plugin/README.md +29 -0
- package/.claude-plugin/marketplace.json +52 -0
- package/.claude-plugin/plugin.json +40 -0
- package/.claude-plugin/settings.json +16 -0
- package/LICENSE +201 -0
- package/README.md +331 -0
- package/WORKFLOW.yaml +36 -0
- package/agents/orchestrator.md +393 -0
- package/commands/forge.md +55 -0
- package/commands/status.md +92 -0
- package/commands/wrap-up.md +85 -0
- package/config.json +17 -0
- package/contexts/init.md +64 -0
- package/contexts/review.md +89 -0
- package/contexts/sprint.md +90 -0
- package/dist/cli/adapters/codex.d.ts +3 -0
- package/dist/cli/adapters/codex.d.ts.map +1 -0
- package/dist/cli/adapters/codex.js +55 -0
- package/dist/cli/adapters/codex.js.map +1 -0
- package/dist/cli/adapters/command-skills.d.ts +5 -0
- package/dist/cli/adapters/command-skills.d.ts.map +1 -0
- package/dist/cli/adapters/command-skills.js +41 -0
- package/dist/cli/adapters/command-skills.js.map +1 -0
- package/dist/cli/adapters/opencode.d.ts +5 -0
- package/dist/cli/adapters/opencode.d.ts.map +1 -0
- package/dist/cli/adapters/opencode.js +41 -0
- package/dist/cli/adapters/opencode.js.map +1 -0
- package/dist/cli/adapters/registry-types.d.ts +13 -0
- package/dist/cli/adapters/registry-types.d.ts.map +1 -0
- package/dist/cli/adapters/registry-types.js +3 -0
- package/dist/cli/adapters/registry-types.js.map +1 -0
- package/dist/cli/adapters/registry.d.ts +7 -0
- package/dist/cli/adapters/registry.d.ts.map +1 -0
- package/dist/cli/adapters/registry.js +46 -0
- package/dist/cli/adapters/registry.js.map +1 -0
- package/dist/cli/app.d.ts +2 -0
- package/dist/cli/app.d.ts.map +1 -0
- package/dist/cli/app.js +46 -0
- package/dist/cli/app.js.map +1 -0
- package/dist/cli/commands/doctor.d.ts +2 -0
- package/dist/cli/commands/doctor.d.ts.map +1 -0
- package/dist/cli/commands/doctor.js +133 -0
- package/dist/cli/commands/doctor.js.map +1 -0
- package/dist/cli/commands/install.d.ts +4 -0
- package/dist/cli/commands/install.d.ts.map +1 -0
- package/dist/cli/commands/install.js +42 -0
- package/dist/cli/commands/install.js.map +1 -0
- package/dist/cli/commands/tui.d.ts +2 -0
- package/dist/cli/commands/tui.d.ts.map +1 -0
- package/dist/cli/commands/tui.js +35 -0
- package/dist/cli/commands/tui.js.map +1 -0
- package/dist/cli/commands/uninstall.d.ts +3 -0
- package/dist/cli/commands/uninstall.d.ts.map +1 -0
- package/dist/cli/commands/uninstall.js +33 -0
- package/dist/cli/commands/uninstall.js.map +1 -0
- package/dist/cli/constants.d.ts +23 -0
- package/dist/cli/constants.d.ts.map +1 -0
- package/dist/cli/constants.js +27 -0
- package/dist/cli/constants.js.map +1 -0
- package/dist/cli/fs.d.ts +12 -0
- package/dist/cli/fs.d.ts.map +1 -0
- package/dist/cli/fs.js +133 -0
- package/dist/cli/fs.js.map +1 -0
- package/dist/cli/help.d.ts +4 -0
- package/dist/cli/help.d.ts.map +1 -0
- package/dist/cli/help.js +28 -0
- package/dist/cli/help.js.map +1 -0
- package/dist/cli/install-plan.d.ts +3 -0
- package/dist/cli/install-plan.d.ts.map +1 -0
- package/dist/cli/install-plan.js +93 -0
- package/dist/cli/install-plan.js.map +1 -0
- package/dist/cli/options.d.ts +8 -0
- package/dist/cli/options.d.ts.map +1 -0
- package/dist/cli/options.js +99 -0
- package/dist/cli/options.js.map +1 -0
- package/dist/cli/state.d.ts +4 -0
- package/dist/cli/state.d.ts.map +1 -0
- package/dist/cli/state.js +13 -0
- package/dist/cli/state.js.map +1 -0
- package/dist/cli/types.d.ts +50 -0
- package/dist/cli/types.d.ts.map +1 -0
- package/dist/cli/types.js +3 -0
- package/dist/cli/types.js.map +1 -0
- package/dist/cli.d.ts +3 -0
- package/dist/cli.d.ts.map +1 -0
- package/dist/cli.js +10 -0
- package/dist/cli.js.map +1 -0
- package/dist/index.d.ts +10 -0
- package/dist/index.d.ts.map +1 -0
- package/dist/index.js +12 -0
- package/dist/index.js.map +1 -0
- package/docs/HOW-TO-USE-CODEX.md +82 -0
- package/docs/HOW-TO-USE-OPENCODE.md +93 -0
- package/docs/agent-adapters.md +152 -0
- package/docs/agents-reference.md +153 -0
- package/docs/architecture.md +159 -0
- package/docs/architecture.mmd +21 -0
- package/docs/cli.md +87 -0
- package/docs/commands-reference.md +138 -0
- package/docs/context-management.md +92 -0
- package/docs/getting-started.md +165 -0
- package/docs/harness-migration.md +95 -0
- package/docs/programmatic-usage.md +109 -0
- package/docs/rules-guide.md +216 -0
- package/package.json +78 -0
- package/rules/context-persistence.md +54 -0
- package/rules/estimation.md +56 -0
- package/rules/learning-rules.md +66 -0
- package/rules/quality-gates.md +65 -0
- package/rules/sprint-discipline.md +49 -0
- package/scripts/check-markdown-links.mjs +97 -0
- package/scripts/check-versions.mjs +46 -0
- package/scripts/make-cli-executable.mjs +12 -0
- package/settings.example.json +27 -0
- package/skills/qa-review/SKILL.md +660 -0
- package/skills/qa-review/manifest.json +41 -0
- package/skills/sprint-forge/SKILL.md +260 -0
- package/skills/sprint-forge/assets/README.md +31 -0
- package/skills/sprint-forge/assets/helpers/analysis-guide.md +207 -0
- package/skills/sprint-forge/assets/helpers/analyzer.md +83 -0
- package/skills/sprint-forge/assets/helpers/debt-tracker.md +122 -0
- package/skills/sprint-forge/assets/helpers/handoff.md +103 -0
- package/skills/sprint-forge/assets/helpers/learner.md +69 -0
- package/skills/sprint-forge/assets/helpers/metrics.md +81 -0
- package/skills/sprint-forge/assets/helpers/reentry-generator.md +121 -0
- package/skills/sprint-forge/assets/helpers/reviewer.md +71 -0
- package/skills/sprint-forge/assets/helpers/sprint-generator.md +145 -0
- package/skills/sprint-forge/assets/modes/INIT.md +204 -0
- package/skills/sprint-forge/assets/modes/SPRINT.md +253 -0
- package/skills/sprint-forge/assets/modes/STATUS.md +145 -0
- package/skills/sprint-forge/assets/templates/PROJECT-README.md +107 -0
- package/skills/sprint-forge/assets/templates/REENTRY-PROMPTS.md +132 -0
- package/skills/sprint-forge/assets/templates/ROADMAP.md +126 -0
- package/skills/sprint-forge/assets/templates/SPRINT.md +189 -0
- package/skills/sprint-forge/manifest.json +165 -0
- package/templates/split-claude-md/AGENTS.md +96 -0
- package/templates/split-claude-md/CLAUDE.md +67 -0
- package/templates/split-claude-md/COMMANDS.md +37 -0
- package/templates/split-claude-md/LEARNED.md +39 -0
- package/templates/split-claude-md/SOUL.md +107 -0
|
@@ -0,0 +1,152 @@
|
|
|
1
|
+
# Agent Adapters
|
|
2
|
+
|
|
3
|
+
Kyro is an agent-agnostic workflow kit. Native platform behavior depends on whether the host agent supports commands, project rules, skills, or plugin manifests. The portable interface is always the same: markdown instructions plus `.agents/kyro/scopes/{scope}/` artifacts.
|
|
4
|
+
|
|
5
|
+
---
|
|
6
|
+
|
|
7
|
+
## Stable Interface
|
|
8
|
+
|
|
9
|
+
Treat these files and directories as Kyro's public interface:
|
|
10
|
+
|
|
11
|
+
| Interface | Purpose |
|
|
12
|
+
|-----------|---------|
|
|
13
|
+
| `agents/orchestrator.md` | Full workflow coordinator instructions |
|
|
14
|
+
| `skills/sprint-forge/SKILL.md` | Sprint planning, execution, status, debt, and re-entry workflow |
|
|
15
|
+
| `skills/qa-review/SKILL.md` | Senior QA, architecture, security, and sprint alignment review |
|
|
16
|
+
| `commands/*.md` | Native slash-command semantics where supported |
|
|
17
|
+
| `.agents/kyro/scopes/{scope}/` | Project roadmap, findings, phases, handoffs, rules, and re-entry prompts |
|
|
18
|
+
|
|
19
|
+
Platforms without slash commands should invoke these equivalent intents:
|
|
20
|
+
|
|
21
|
+
| Intent | Equivalent request |
|
|
22
|
+
|--------|--------------------|
|
|
23
|
+
| `forge` | Analyze, plan, execute, review, and close the sprint |
|
|
24
|
+
| `status` | Read artifacts and report project progress/debt |
|
|
25
|
+
| `wrap-up` | Close the session and update re-entry prompts |
|
|
26
|
+
|
|
27
|
+
---
|
|
28
|
+
|
|
29
|
+
## Generic Setup
|
|
30
|
+
|
|
31
|
+
Copy or symlink Kyro into the target project:
|
|
32
|
+
|
|
33
|
+
```bash
|
|
34
|
+
mkdir -p .skills .agents .agents/kyro/scopes
|
|
35
|
+
|
|
36
|
+
cp -r /path/to/kyro-ai/skills/sprint-forge .skills/
|
|
37
|
+
cp -r /path/to/kyro-ai/skills/qa-review .skills/
|
|
38
|
+
cp /path/to/kyro-ai/agents/orchestrator.md .agents/
|
|
39
|
+
```
|
|
40
|
+
|
|
41
|
+
Use this onboarding prompt for any agent:
|
|
42
|
+
|
|
43
|
+
```text
|
|
44
|
+
Use Kyro as the workflow for this project.
|
|
45
|
+
|
|
46
|
+
Read these files first:
|
|
47
|
+
- .agents/orchestrator.md
|
|
48
|
+
- .skills/sprint-forge/SKILL.md
|
|
49
|
+
- .skills/qa-review/SKILL.md
|
|
50
|
+
|
|
51
|
+
Persist workflow artifacts under:
|
|
52
|
+
- .agents/kyro/scopes/{scope}/
|
|
53
|
+
|
|
54
|
+
If native slash commands are unavailable:
|
|
55
|
+
- forge = analyze/plan/execute/review/close
|
|
56
|
+
- status = read artifacts and report progress/debt
|
|
57
|
+
- wrap-up = close session and update re-entry prompts
|
|
58
|
+
```
|
|
59
|
+
|
|
60
|
+
---
|
|
61
|
+
|
|
62
|
+
## Claude Code Adapter
|
|
63
|
+
|
|
64
|
+
Claude Code has a native adapter through `.claude-plugin/`.
|
|
65
|
+
|
|
66
|
+
```bash
|
|
67
|
+
/plugin marketplace add SynapSync/kyro-ai
|
|
68
|
+
/plugin install kyro-ai@kyro-ai
|
|
69
|
+
```
|
|
70
|
+
|
|
71
|
+
The Claude adapter registers commands, the orchestrator agent, and skills. It is the only native adapter included in this repository today.
|
|
72
|
+
|
|
73
|
+
---
|
|
74
|
+
|
|
75
|
+
## Codex Adapter
|
|
76
|
+
|
|
77
|
+
Codex-style agents should use Kyro as project context:
|
|
78
|
+
|
|
79
|
+
```bash
|
|
80
|
+
mkdir -p .skills .agents
|
|
81
|
+
cp -r kyro-ai/skills/sprint-forge .skills/
|
|
82
|
+
cp -r kyro-ai/skills/qa-review .skills/
|
|
83
|
+
cp kyro-ai/agents/orchestrator.md .agents/
|
|
84
|
+
```
|
|
85
|
+
|
|
86
|
+
Prompt:
|
|
87
|
+
|
|
88
|
+
```text
|
|
89
|
+
Read .agents/orchestrator.md and the Kyro skills in .skills/.
|
|
90
|
+
Use the forge intent for this scope: {scope}.
|
|
91
|
+
Persist outputs under .agents/kyro/scopes/{scope}/.
|
|
92
|
+
```
|
|
93
|
+
|
|
94
|
+
Native command registration depends on the Codex environment. If slash commands are unavailable, use the manual intent names.
|
|
95
|
+
|
|
96
|
+
---
|
|
97
|
+
|
|
98
|
+
## OpenCode Adapter
|
|
99
|
+
|
|
100
|
+
OpenCode usage is manual unless your environment supports project-level rule files.
|
|
101
|
+
|
|
102
|
+
```bash
|
|
103
|
+
mkdir -p .skills .agents
|
|
104
|
+
cp -r kyro-ai/skills/sprint-forge .skills/
|
|
105
|
+
cp -r kyro-ai/skills/qa-review .skills/
|
|
106
|
+
cp kyro-ai/agents/orchestrator.md .agents/
|
|
107
|
+
```
|
|
108
|
+
|
|
109
|
+
Reference the files in the AI panel:
|
|
110
|
+
|
|
111
|
+
```text
|
|
112
|
+
@file .agents/orchestrator.md
|
|
113
|
+
@file .skills/sprint-forge/SKILL.md
|
|
114
|
+
@file .skills/qa-review/SKILL.md
|
|
115
|
+
|
|
116
|
+
Run the status intent for .agents/kyro/scopes/{scope}/.
|
|
117
|
+
```
|
|
118
|
+
|
|
119
|
+
---
|
|
120
|
+
|
|
121
|
+
## Cursor Adapter
|
|
122
|
+
|
|
123
|
+
Cursor can use Kyro through project rules and referenced files.
|
|
124
|
+
|
|
125
|
+
Recommended setup:
|
|
126
|
+
|
|
127
|
+
1. Copy the Kyro files using the generic setup.
|
|
128
|
+
2. Add a Cursor project rule that tells the agent to read `.agents/orchestrator.md`.
|
|
129
|
+
3. Ask Cursor to persist sprint artifacts under `.agents/kyro/scopes/{scope}/`.
|
|
130
|
+
|
|
131
|
+
Cursor prompt:
|
|
132
|
+
|
|
133
|
+
```text
|
|
134
|
+
Use Kyro for this task. Read .agents/orchestrator.md and the skills under .skills/.
|
|
135
|
+
Run the forge intent for {scope}. Do not create a custom planning format.
|
|
136
|
+
```
|
|
137
|
+
|
|
138
|
+
---
|
|
139
|
+
|
|
140
|
+
## Model Guidance
|
|
141
|
+
|
|
142
|
+
Kyro does not route or enforce model selection.
|
|
143
|
+
|
|
144
|
+
- Use the strongest available model for implementation, debugging, and architecture decisions.
|
|
145
|
+
- Lighter/faster models are acceptable for read-only analysis, status reports, and documentation review.
|
|
146
|
+
- When the platform supports model overrides, choose per task rather than encoding model names into Kyro artifacts.
|
|
147
|
+
|
|
148
|
+
---
|
|
149
|
+
|
|
150
|
+
## Compatibility Rule
|
|
151
|
+
|
|
152
|
+
Do not add platform-specific behavior to the core workflow unless the behavior works through markdown artifacts or is isolated in an adapter-specific directory.
|
|
@@ -0,0 +1,153 @@
|
|
|
1
|
+
# Agents Reference
|
|
2
|
+
|
|
3
|
+
Kyro uses one agent: the **orchestrator**. It coordinates the sprint lifecycle and owns analysis, planning, implementation, review, debugging, and checkpoint behavior.
|
|
4
|
+
|
|
5
|
+
---
|
|
6
|
+
|
|
7
|
+
## Overview
|
|
8
|
+
|
|
9
|
+
| Agent | Role | Model | Tools | Memory |
|
|
10
|
+
|-------|------|-------|-------|--------|
|
|
11
|
+
| **orchestrator** | Full cycle coordination, validation gates, and checkpoints | opus | Read, Glob, Grep, Bash, Edit, Write | project |
|
|
12
|
+
|
|
13
|
+
---
|
|
14
|
+
|
|
15
|
+
## Orchestrator
|
|
16
|
+
|
|
17
|
+
**File:** `agents/orchestrator.md`
|
|
18
|
+
|
|
19
|
+
The orchestrator coordinates the complete sprint lifecycle. It is the brain of `/kyro:forge`, manages gates, executes protocols, runs checkpoints, and handles sprint close.
|
|
20
|
+
|
|
21
|
+
### When Triggered
|
|
22
|
+
|
|
23
|
+
- `/kyro:forge` command.
|
|
24
|
+
- Any workflow task requiring coordinated analysis, planning, implementation, or review.
|
|
25
|
+
|
|
26
|
+
### Tools
|
|
27
|
+
|
|
28
|
+
| Tool | Usage |
|
|
29
|
+
|------|-------|
|
|
30
|
+
| Read | Read sprint files, roadmaps, rules, and code paths |
|
|
31
|
+
| Glob | Find project files by pattern |
|
|
32
|
+
| Grep | Search codebase for patterns, debug artifacts, secrets |
|
|
33
|
+
| Bash | Run commands, tests, quality gates, and read-only analysis |
|
|
34
|
+
| Edit | Modify code files during implementation |
|
|
35
|
+
| Write | Create sprint documents, findings, roadmaps, and re-entry prompts |
|
|
36
|
+
|
|
37
|
+
---
|
|
38
|
+
|
|
39
|
+
## Protocols
|
|
40
|
+
|
|
41
|
+
### Analysis Protocol
|
|
42
|
+
|
|
43
|
+
Used during INIT and codebase exploration. The orchestrator performs read-only analysis:
|
|
44
|
+
|
|
45
|
+
1. Detect work type: audit/refactor, new feature, bugfix, new project, or tech debt.
|
|
46
|
+
2. Explore architecture, code quality, dependencies, risks, and visible debt.
|
|
47
|
+
3. Generate findings and recommendations for the roadmap.
|
|
48
|
+
|
|
49
|
+
During analysis, the orchestrator must not edit files.
|
|
50
|
+
|
|
51
|
+
### Review Checklist
|
|
52
|
+
|
|
53
|
+
Used after each task completion during sprint execution.
|
|
54
|
+
|
|
55
|
+
**BLOCKER**
|
|
56
|
+
|
|
57
|
+
- Related tests pass when the project defines tests.
|
|
58
|
+
- Typecheck/build passes.
|
|
59
|
+
- No debug artifacts are left in production code.
|
|
60
|
+
- No hardcoded secrets or credentials.
|
|
61
|
+
- No syntax errors or broken imports.
|
|
62
|
+
|
|
63
|
+
**WARNING**
|
|
64
|
+
|
|
65
|
+
- New code has appropriate coverage when practical.
|
|
66
|
+
- Non-obvious logic is documented.
|
|
67
|
+
- New debt is tracked.
|
|
68
|
+
- No visible performance regressions.
|
|
69
|
+
|
|
70
|
+
**SUGGESTION**
|
|
71
|
+
|
|
72
|
+
- Code follows project conventions.
|
|
73
|
+
- Refactoring opportunities are noted for retro.
|
|
74
|
+
- Related documentation updates are identified.
|
|
75
|
+
|
|
76
|
+
### Debug Protocol
|
|
77
|
+
|
|
78
|
+
Used when a task fails validation or runtime behavior breaks:
|
|
79
|
+
|
|
80
|
+
1. Reproduce the failure.
|
|
81
|
+
2. Generate ranked hypotheses.
|
|
82
|
+
3. Investigate evidence.
|
|
83
|
+
4. Identify root cause.
|
|
84
|
+
5. Propose a fix.
|
|
85
|
+
6. Escalate if unresolved after three investigation rounds.
|
|
86
|
+
|
|
87
|
+
The orchestrator waits for approval before applying non-trivial fixes.
|
|
88
|
+
|
|
89
|
+
---
|
|
90
|
+
|
|
91
|
+
## Checkpoint Protocol
|
|
92
|
+
|
|
93
|
+
The orchestrator owns checkpoints directly. They are not delegated to a separate agent.
|
|
94
|
+
|
|
95
|
+
| Checkpoint | When | Purpose |
|
|
96
|
+
|------------|------|---------|
|
|
97
|
+
| startup | Start of orchestration | Load rules, detect active sprint, summarize state |
|
|
98
|
+
| pre-phase | Before each phase | Validate rules, sprint file, and worktree state |
|
|
99
|
+
| rule check | Before task execution | Detect likely violations of learned rules |
|
|
100
|
+
| post-edit scan | After edits | Search changed files for debug artifacts and secrets |
|
|
101
|
+
| task complete | After task validation | Verify task status, checkpoint, remaining work, and new debt |
|
|
102
|
+
| pre-commit | Before commit | Run quality gates and final post-edit scan |
|
|
103
|
+
| learn capture | Sprint close | Propose new rules from corrections and discoveries |
|
|
104
|
+
|
|
105
|
+
---
|
|
106
|
+
|
|
107
|
+
## Gate Protocol
|
|
108
|
+
|
|
109
|
+
At each gate, the orchestrator presents:
|
|
110
|
+
|
|
111
|
+
```text
|
|
112
|
+
===================================
|
|
113
|
+
GATE [N]: [Phase Name] Complete
|
|
114
|
+
===================================
|
|
115
|
+
|
|
116
|
+
Summary:
|
|
117
|
+
- [key outcomes from this phase]
|
|
118
|
+
|
|
119
|
+
Next phase: [what happens next]
|
|
120
|
+
|
|
121
|
+
Options:
|
|
122
|
+
-> "proceed" -- continue to next phase
|
|
123
|
+
-> "adjust" -- modify before continuing
|
|
124
|
+
-> "cancel" -- stop the workflow
|
|
125
|
+
```
|
|
126
|
+
|
|
127
|
+
The orchestrator never proceeds past a gate without explicit user approval.
|
|
128
|
+
|
|
129
|
+
---
|
|
130
|
+
|
|
131
|
+
## Sprint Close Protocol
|
|
132
|
+
|
|
133
|
+
After all tasks are complete:
|
|
134
|
+
|
|
135
|
+
1. Run the pre-commit checkpoint.
|
|
136
|
+
2. Consolidate findings.
|
|
137
|
+
3. Fill the retrospective.
|
|
138
|
+
4. Update accumulated technical debt in markdown.
|
|
139
|
+
5. Update frontmatter.
|
|
140
|
+
6. Generate or update re-entry prompts.
|
|
141
|
+
7. Update roadmap if execution changed the plan.
|
|
142
|
+
8. Run the learn-capture checkpoint.
|
|
143
|
+
9. Propose new rules for `.agents/kyro/scopes/rules.md`.
|
|
144
|
+
|
|
145
|
+
---
|
|
146
|
+
|
|
147
|
+
## Constraints
|
|
148
|
+
|
|
149
|
+
- Never skip phases or gates.
|
|
150
|
+
- Never proceed without user approval at gates.
|
|
151
|
+
- If implementation reveals the plan was wrong, return to planning.
|
|
152
|
+
- Capture learnings and propose rules at the end of every cycle.
|
|
153
|
+
- Keep the user informed at each step.
|
|
@@ -0,0 +1,159 @@
|
|
|
1
|
+
# Architecture
|
|
2
|
+
|
|
3
|
+
This document describes Kyro's Command > Agent > Skill workflow architecture and the markdown artifacts used to preserve project context.
|
|
4
|
+
|
|
5
|
+
---
|
|
6
|
+
|
|
7
|
+
## Command > Agent > Skill Pattern
|
|
8
|
+
|
|
9
|
+
Kyro is organized in three layers:
|
|
10
|
+
|
|
11
|
+
```
|
|
12
|
+
User Command (/kyro:forge, /kyro:status, /kyro:wrap-up)
|
|
13
|
+
|
|
|
14
|
+
v
|
|
15
|
+
Agent (orchestrator)
|
|
16
|
+
|
|
|
17
|
+
+---> Skill (core)
|
|
18
|
+
|
|
|
19
|
+
+---> Skill (qa-review)
|
|
20
|
+
```
|
|
21
|
+
|
|
22
|
+
### Commands
|
|
23
|
+
|
|
24
|
+
Commands are the user-facing interface. Each command is defined as a markdown file in `commands/` with frontmatter that specifies its description and argument hints.
|
|
25
|
+
|
|
26
|
+
| Command | Primary Agent | Purpose |
|
|
27
|
+
|---------|--------------|---------|
|
|
28
|
+
| `/kyro:forge` | orchestrator | Full cycle: Analyze, Plan, Implement, Review, Close |
|
|
29
|
+
| `/kyro:status` | orchestrator | Read-only project progress and debt summary |
|
|
30
|
+
| `/kyro:wrap-up` | orchestrator | End-of-session closure ritual with quality check and context handoff |
|
|
31
|
+
|
|
32
|
+
### Agent
|
|
33
|
+
|
|
34
|
+
The orchestrator coordinates the full sprint lifecycle. It performs read-only analysis during discovery, generates plans, executes approved tasks, runs validation, handles debugging, updates sprint artifacts, and owns lifecycle checkpoints.
|
|
35
|
+
|
|
36
|
+
| Agent | Toolset | Can Write? | Role |
|
|
37
|
+
|-------|---------|-----------|------|
|
|
38
|
+
| orchestrator | Read, Glob, Grep, Bash, Edit, Write | Yes | Full lifecycle coordination |
|
|
39
|
+
|
|
40
|
+
### Skills
|
|
41
|
+
|
|
42
|
+
Skills provide domain knowledge that the orchestrator consumes.
|
|
43
|
+
|
|
44
|
+
| Skill | Knowledge Domain |
|
|
45
|
+
|-------|-----------------|
|
|
46
|
+
| `core` | Core orchestration: modes, helpers, templates, gates, re-entry prompts |
|
|
47
|
+
| `qa-review` | Senior QA audit, architecture validation, security review, sprint alignment |
|
|
48
|
+
|
|
49
|
+
---
|
|
50
|
+
|
|
51
|
+
## Data Flow
|
|
52
|
+
|
|
53
|
+
```
|
|
54
|
+
USER
|
|
55
|
+
|
|
|
56
|
+
v
|
|
57
|
+
/kyro:forge
|
|
58
|
+
|
|
|
59
|
+
v
|
|
60
|
+
ORCHESTRATOR
|
|
61
|
+
|-- loads .agents/kyro/scopes/rules.md
|
|
62
|
+
|-- reads sprint-forge skill assets
|
|
63
|
+
|-- runs built-in checkpoints
|
|
64
|
+
|
|
|
65
|
+
v
|
|
66
|
+
.agents/kyro/scopes/{scope}/
|
|
67
|
+
|-- findings/
|
|
68
|
+
|-- phases/
|
|
69
|
+
|-- ROADMAP.md
|
|
70
|
+
|-- README.md
|
|
71
|
+
|-- RE-ENTRY-PROMPTS.md
|
|
72
|
+
```
|
|
73
|
+
|
|
74
|
+
### Flow for `/kyro:forge`
|
|
75
|
+
|
|
76
|
+
1. **Rules Loading** - Orchestrator reads `.agents/kyro/scopes/rules.md` if present.
|
|
77
|
+
2. **Analysis** - Orchestrator explores the codebase and creates finding files.
|
|
78
|
+
3. **Gate 1** - User approves analysis.
|
|
79
|
+
4. **Planning** - Orchestrator generates a sprint document with phases and tasks.
|
|
80
|
+
5. **Gate 2** - User approves the plan.
|
|
81
|
+
6. **Implementation** - Orchestrator executes tasks, runs review checks, and checkpoints after each phase.
|
|
82
|
+
7. **Gate 3** - User approves implementation.
|
|
83
|
+
8. **Review and Close** - Orchestrator runs retro, updates debt tables in markdown, proposes rules, and updates re-entry prompts.
|
|
84
|
+
|
|
85
|
+
---
|
|
86
|
+
|
|
87
|
+
## Artifact Layout
|
|
88
|
+
|
|
89
|
+
Kyro stores workflow state in markdown files. This keeps the workflow portable across AI coding platforms, easy to review in git, and usable without a local service.
|
|
90
|
+
|
|
91
|
+
```
|
|
92
|
+
.agents/kyro/scopes/
|
|
93
|
+
├── rules.md
|
|
94
|
+
└── {scope}/
|
|
95
|
+
├── README.md
|
|
96
|
+
├── ROADMAP.md
|
|
97
|
+
├── RE-ENTRY-PROMPTS.md
|
|
98
|
+
├── findings/
|
|
99
|
+
├── phases/
|
|
100
|
+
└── handoffs/
|
|
101
|
+
```
|
|
102
|
+
|
|
103
|
+
`{scope}` is the work topic in kebab-case, for example `oauth-implementation` or `ui-redesign`.
|
|
104
|
+
|
|
105
|
+
The output directory path (`{output_kyro_dir}`) is resolved once at the start of any mode and embedded in `README.md` and `RE-ENTRY-PROMPTS.md`. These two files are the source of truth for the path.
|
|
106
|
+
|
|
107
|
+
---
|
|
108
|
+
|
|
109
|
+
## Built-In Checkpoints
|
|
110
|
+
|
|
111
|
+
The orchestrator runs checkpoints at lifecycle moments:
|
|
112
|
+
|
|
113
|
+
| Checkpoint | Purpose |
|
|
114
|
+
|------------|---------|
|
|
115
|
+
| startup | Load rules and detect active sprint state |
|
|
116
|
+
| pre-phase | Validate state before a phase starts |
|
|
117
|
+
| rule check | Verify relevant learned rules |
|
|
118
|
+
| post-edit scan | Detect debug artifacts and likely secrets |
|
|
119
|
+
| task complete | Verify task status and checkpoint state |
|
|
120
|
+
| pre-commit | Run configured quality gates |
|
|
121
|
+
| learn capture | Propose new rules from corrections |
|
|
122
|
+
|
|
123
|
+
---
|
|
124
|
+
|
|
125
|
+
## How the Workflow Differs from v1.x
|
|
126
|
+
|
|
127
|
+
Kyro v2.0 is a full workflow that replaces the v1.x single-skill approach.
|
|
128
|
+
|
|
129
|
+
```
|
|
130
|
+
v1.x: User message -> sprint-forge skill -> markdown artifacts
|
|
131
|
+
|
|
132
|
+
v2.0: User command -> orchestrator
|
|
133
|
+
-> core / qa-review skills
|
|
134
|
+
-> built-in checkpoints
|
|
135
|
+
-> markdown artifacts
|
|
136
|
+
```
|
|
137
|
+
|
|
138
|
+
| Dimension | v1.x | v2.0 |
|
|
139
|
+
|-----------|------|------|
|
|
140
|
+
| Type | Single skill | Full workflow with commands, one agent, skills, and checkpoints |
|
|
141
|
+
| Entry point | Text triggers | Slash commands |
|
|
142
|
+
| Learning | Per-project retro only | Persistent rules in `.agents/kyro/scopes/rules.md` |
|
|
143
|
+
| Agent | Skill-only execution | Orchestrator |
|
|
144
|
+
| Quality gates | Basic | Per-task checklist + approval gates |
|
|
145
|
+
| Context transfer | Re-entry prompts | Re-entry prompts + enriched handoffs |
|
|
146
|
+
| State model | Markdown artifacts | Markdown artifacts |
|
|
147
|
+
|
|
148
|
+
---
|
|
149
|
+
|
|
150
|
+
## Component Map
|
|
151
|
+
|
|
152
|
+
| Component | Location |
|
|
153
|
+
|-----------|----------|
|
|
154
|
+
| Commands | `commands/` |
|
|
155
|
+
| Orchestrator | `agents/orchestrator.md` |
|
|
156
|
+
| Sprint workflow skill | `skills/sprint-forge/` |
|
|
157
|
+
| QA review skill | `skills/qa-review/` |
|
|
158
|
+
| Templates | `skills/sprint-forge/assets/templates/` |
|
|
159
|
+
| Rules | `.agents/kyro/scopes/rules.md` in the target project |
|
|
@@ -0,0 +1,21 @@
|
|
|
1
|
+
flowchart TD
|
|
2
|
+
USER["User command"]
|
|
3
|
+
FORGE["/kyro:forge"]
|
|
4
|
+
STATUS["/kyro:status"]
|
|
5
|
+
WRAP["/kyro:wrap-up"]
|
|
6
|
+
ORCH["orchestrator agent"]
|
|
7
|
+
SPRINT["sprint-forge skill"]
|
|
8
|
+
QA["qa-review skill"]
|
|
9
|
+
RULES[".agents/kyro/scopes/rules.md"]
|
|
10
|
+
SCOPE[".agents/kyro/scopes/{scope}/ markdown artifacts"]
|
|
11
|
+
|
|
12
|
+
USER --> FORGE
|
|
13
|
+
USER --> STATUS
|
|
14
|
+
USER --> WRAP
|
|
15
|
+
FORGE --> ORCH
|
|
16
|
+
WRAP --> ORCH
|
|
17
|
+
STATUS --> SCOPE
|
|
18
|
+
ORCH --> SPRINT
|
|
19
|
+
ORCH --> QA
|
|
20
|
+
ORCH --> RULES
|
|
21
|
+
ORCH --> SCOPE
|
package/docs/cli.md
ADDED
|
@@ -0,0 +1,87 @@
|
|
|
1
|
+
# Kyro CLI
|
|
2
|
+
|
|
3
|
+
Kyro includes a small CLI for installing workspace harness assets and checking health.
|
|
4
|
+
|
|
5
|
+
## Commands
|
|
6
|
+
|
|
7
|
+
```bash
|
|
8
|
+
kyro # Open the interactive TUI
|
|
9
|
+
kyro install # Install adapter assets, OpenCode workspace by default
|
|
10
|
+
kyro doctor # Read-only package/workspace health check
|
|
11
|
+
kyro sync # Refresh managed workspace assets
|
|
12
|
+
kyro uninstall # Remove managed workspace assets, preserving Kyro project state
|
|
13
|
+
```
|
|
14
|
+
|
|
15
|
+
`npx kyro-ai` also resolves to the same CLI entrypoint.
|
|
16
|
+
|
|
17
|
+
## Install Scope
|
|
18
|
+
|
|
19
|
+
The default install scope is `workspace`.
|
|
20
|
+
|
|
21
|
+
Workspace install writes project-local files so the repository carries its Kyro harness configuration:
|
|
22
|
+
|
|
23
|
+
```text
|
|
24
|
+
.agents/
|
|
25
|
+
├── kyro-ai/
|
|
26
|
+
│ ├── sprint-forge/
|
|
27
|
+
│ ├── commands/
|
|
28
|
+
│ ├── skills/
|
|
29
|
+
│ ├── KYRO.md
|
|
30
|
+
│ └── manifest.json
|
|
31
|
+
├── skills/
|
|
32
|
+
│ ├── kyro-forge/SKILL.md
|
|
33
|
+
│ ├── kyro-status/SKILL.md
|
|
34
|
+
│ └── kyro-wrap-up/SKILL.md
|
|
35
|
+
└── core/
|
|
36
|
+
└── kyro.json
|
|
37
|
+
```
|
|
38
|
+
|
|
39
|
+
## OpenCode Adapter
|
|
40
|
+
|
|
41
|
+
OpenCode and Codex are the workspace adapters implemented by the CLI:
|
|
42
|
+
|
|
43
|
+
```bash
|
|
44
|
+
kyro install --agent opencode --scope workspace --dry-run
|
|
45
|
+
kyro install --agent opencode --scope workspace --yes
|
|
46
|
+
kyro install --agent codex --scope workspace --yes
|
|
47
|
+
```
|
|
48
|
+
|
|
49
|
+
The adapters project Kyro workflows into `.agents/skills/` so compatible agents can discover command-like skills without asking the user to invoke Kyro through prose.
|
|
50
|
+
|
|
51
|
+
Projected skills:
|
|
52
|
+
|
|
53
|
+
- `kyro-forge`
|
|
54
|
+
- `kyro-status`
|
|
55
|
+
- `kyro-wrap-up`
|
|
56
|
+
|
|
57
|
+
Each projected skill references the managed Kyro core in `.agents/kyro/internal/` instead of duplicating long workflow instructions.
|
|
58
|
+
|
|
59
|
+
## State Model
|
|
60
|
+
|
|
61
|
+
`kyro install` creates only global project state:
|
|
62
|
+
|
|
63
|
+
```text
|
|
64
|
+
.agents/kyro/scopes/kyro.json
|
|
65
|
+
```
|
|
66
|
+
|
|
67
|
+
It does not create scoped state. Scoped state is created later when a scope is created or opened by a future scope/forge workflow.
|
|
68
|
+
|
|
69
|
+
Initial state shape:
|
|
70
|
+
|
|
71
|
+
```json
|
|
72
|
+
{
|
|
73
|
+
"schemaVersion": 1,
|
|
74
|
+
"artifactRoot": ".agents/kyro/scopes",
|
|
75
|
+
"scopes": [],
|
|
76
|
+
"activeScope": null,
|
|
77
|
+
"installedAdapters": []
|
|
78
|
+
}
|
|
79
|
+
```
|
|
80
|
+
|
|
81
|
+
## Claude Plugin Support
|
|
82
|
+
|
|
83
|
+
The Claude plugin adapter remains first-class through `.claude-plugin/`. The CLI does not replace it; it complements Kyro's adapter story for agents that need workspace-installed commands, skills, root AGENTS.md managed blocks, and core assets.
|
|
84
|
+
|
|
85
|
+
## Unsupported Generic Adapter
|
|
86
|
+
|
|
87
|
+
Kyro does not provide `--agent generic`. Cross-agent instructions belong in the root `AGENTS.md` standard, and adapter installs should target concrete agent capabilities.
|
|
@@ -0,0 +1,138 @@
|
|
|
1
|
+
# Commands Reference
|
|
2
|
+
|
|
3
|
+
Kyro provides 3 slash commands. Each command maps to one or more agents and skills that handle the actual work.
|
|
4
|
+
|
|
5
|
+
---
|
|
6
|
+
|
|
7
|
+
## /kyro:forge
|
|
8
|
+
|
|
9
|
+
**Full sprint cycle: Analyze, Plan, Implement, Review, Close.**
|
|
10
|
+
|
|
11
|
+
### Syntax
|
|
12
|
+
|
|
13
|
+
```
|
|
14
|
+
/kyro:forge <project path or description>
|
|
15
|
+
```
|
|
16
|
+
|
|
17
|
+
### Arguments
|
|
18
|
+
|
|
19
|
+
The argument describes what to analyze or work on. It can be a path, a module name, or a description of the work.
|
|
20
|
+
|
|
21
|
+
### Examples
|
|
22
|
+
|
|
23
|
+
```
|
|
24
|
+
/kyro:forge analyze the authentication module
|
|
25
|
+
/kyro:forge audit code quality in src/api/
|
|
26
|
+
/kyro:forge refactor the persistence layer
|
|
27
|
+
/kyro:forge add user profile feature
|
|
28
|
+
/kyro:forge fix the login timeout bug
|
|
29
|
+
```
|
|
30
|
+
|
|
31
|
+
### Phases and Gates
|
|
32
|
+
|
|
33
|
+
The `/kyro:forge` command runs the complete lifecycle:
|
|
34
|
+
|
|
35
|
+
```
|
|
36
|
+
[GATE 0: RULES] Load learned rules from .agents/kyro/scopes/rules.md
|
|
37
|
+
|
|
|
38
|
+
[PHASE 1: ANALYZE] Analysis phase investigates codebase (read-only)
|
|
39
|
+
|
|
|
40
|
+
GATE 1 User approves analysis and plan direction
|
|
41
|
+
|
|
|
42
|
+
[PHASE 2: PLAN] Generate sprint with phases, tasks, and estimates
|
|
43
|
+
|
|
|
44
|
+
GATE 2 User approves sprint plan
|
|
45
|
+
|
|
|
46
|
+
[PHASE 3: IMPLEMENT] Execute task by task
|
|
47
|
+
|-- After each task: Review step validates (BLOCKER/WARNING/SUGGESTION)
|
|
48
|
+
|-- On failure: Debug protocol investigates root cause
|
|
49
|
+
|-- After each phase: Checkpoint saved to disk
|
|
50
|
+
|
|
|
51
|
+
GATE 3 User approves implementation
|
|
52
|
+
|
|
|
53
|
+
[PHASE 4: REVIEW] Full sprint review + retrospective
|
|
54
|
+
|
|
|
55
|
+
[PHASE 5: CLOSE] Debt update, re-entry prompts, rule proposals
|
|
56
|
+
```
|
|
57
|
+
|
|
58
|
+
### Gate Options
|
|
59
|
+
|
|
60
|
+
At each gate, the orchestrator presents a summary and waits for your decision:
|
|
61
|
+
|
|
62
|
+
| Option | Effect |
|
|
63
|
+
|--------|--------|
|
|
64
|
+
| `proceed` | Continue to the next phase |
|
|
65
|
+
| `adjust` | Modify the output before continuing (describe what to change) |
|
|
66
|
+
| `cancel` | Stop the workflow |
|
|
67
|
+
|
|
68
|
+
### Orchestrator Protocols
|
|
69
|
+
|
|
70
|
+
- **Analysis protocol** -- Phase 1 (codebase exploration, read-only)
|
|
71
|
+
- **Review checklist** -- Phase 3 (after each task)
|
|
72
|
+
- **Debug protocol** -- Phase 3 (on task failure)
|
|
73
|
+
- **orchestrator** -- coordinates the full cycle
|
|
74
|
+
|
|
75
|
+
---
|
|
76
|
+
|
|
77
|
+
## /kyro:status
|
|
78
|
+
|
|
79
|
+
**Project progress, sprint state, and technical debt summary.**
|
|
80
|
+
|
|
81
|
+
### Syntax
|
|
82
|
+
|
|
83
|
+
```
|
|
84
|
+
/kyro:status [brief|full|debt]
|
|
85
|
+
```
|
|
86
|
+
|
|
87
|
+
### Variants
|
|
88
|
+
|
|
89
|
+
| Variant | What It Shows |
|
|
90
|
+
|---------|---------------|
|
|
91
|
+
| `brief` | Sprint progress bars and next sprint preview only |
|
|
92
|
+
| `full` | Complete report with all sections (default) |
|
|
93
|
+
| `debt` | Technical debt table and aged debt items |
|
|
94
|
+
|
|
95
|
+
### Examples
|
|
96
|
+
|
|
97
|
+
```
|
|
98
|
+
/kyro:status # Full report
|
|
99
|
+
/kyro:status brief # Quick progress check
|
|
100
|
+
/kyro:status debt # Focus on technical debt
|
|
101
|
+
```
|
|
102
|
+
|
|
103
|
+
### Report Sections
|
|
104
|
+
|
|
105
|
+
The full report includes:
|
|
106
|
+
|
|
107
|
+
```
|
|
108
|
+
KYRO -- Project Status
|
|
109
|
+
|
|
110
|
+
## Sprint Progress
|
|
111
|
+
Sprint 1: xxxxxxxxxx 10/10 (100%) Complete
|
|
112
|
+
Sprint 2: xxxxxxxx-- 8/10 ( 80%) Complete
|
|
113
|
+
Sprint 3: xxxxxxx--- 7/10 ( 70%) In Progress
|
|
114
|
+
|
|
115
|
+
## Technical Debt
|
|
116
|
+
- Open: 4
|
|
117
|
+
- In progress: 1
|
|
118
|
+
- Aged: 2
|
|
119
|
+
- Critical: 1
|
|
120
|
+
|
|
121
|
+
## Roadmap Health
|
|
122
|
+
- Sprints completed: 2/5
|
|
123
|
+
- Roadmap adaptations: 1
|
|
124
|
+
- Carry-over tasks: 3
|
|
125
|
+
|
|
126
|
+
## Next Sprint Preview
|
|
127
|
+
Sprint 4: [title]
|
|
128
|
+
- Suggested phases: [count]
|
|
129
|
+
- Carry-over tasks: [count]
|
|
130
|
+
- Critical debt items due: [count]
|
|
131
|
+
```
|
|
132
|
+
|
|
133
|
+
### Data Sources
|
|
134
|
+
|
|
135
|
+
The status command reads all files in the output directory:
|
|
136
|
+
- `README.md` for project overview
|
|
137
|
+
- `ROADMAP.md` for planned sprints
|
|
138
|
+
- All `phases/SPRINT-*.md` files for progress, debt, and retro data
|