loki-mode 6.42.0 → 6.44.0
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/README.md +206 -141
- package/SKILL.md +2 -2
- package/VERSION +1 -1
- package/autonomy/loki +20 -3
- package/dashboard/__init__.py +1 -1
- package/docs/INSTALLATION.md +1 -1
- package/package.json +1 -1
- package/web-app/dist/assets/{Badge-Cv068X_w.js → Badge-aPxpGmfn.js} +1 -1
- package/web-app/dist/assets/{Button-DjxnqObR.js → Button-DlY3TKRG.js} +1 -1
- package/web-app/dist/assets/{Card-Ct8CoaX9.js → Card-DvU1jTxb.js} +1 -1
- package/web-app/dist/assets/HomePage-BYOHEkPT.js +18 -0
- package/web-app/dist/assets/ProjectPage-BxnUaUte.js +141 -0
- package/web-app/dist/assets/{ProjectsPage-uAcgLclp.js → ProjectsPage-BNbxt9RO.js} +1 -1
- package/web-app/dist/assets/{SettingsPage-CD0Vmtdu.js → SettingsPage-DXhzzC5L.js} +1 -1
- package/web-app/dist/assets/{TemplatesPage-CuWAlBEX.js → TemplatesPage-DwSAg4jO.js} +1 -1
- package/web-app/dist/assets/TerminalOutput-D9l4hRcp.js +31 -0
- package/web-app/dist/assets/{clock-BZGjjvpN.js → clock-KRe2srsu.js} +1 -1
- package/web-app/dist/assets/{external-link-DMhX_uRl.js → external-link-BSjKSqVI.js} +1 -1
- package/web-app/dist/assets/{index-ABzvlaPu.js → index-DKbfn_6W.js} +2 -2
- package/web-app/dist/assets/index-Dn3MgXE4.css +1 -0
- package/web-app/dist/index.html +2 -2
- package/web-app/dist/assets/HomePage-CwHuVFzN.js +0 -33
- package/web-app/dist/assets/ProjectPage-CCzfXsfH.js +0 -116
- package/web-app/dist/assets/TerminalOutput-jc4B3JYH.js +0 -16
- package/web-app/dist/assets/index-C-IWanPu.css +0 -1
package/README.md
CHANGED
|
@@ -1,163 +1,176 @@
|
|
|
1
|
+
<div align="center">
|
|
2
|
+
|
|
1
3
|
# Loki Mode
|
|
2
4
|
|
|
3
|
-
|
|
5
|
+
### Autonomous multi-agent development with self-verification
|
|
4
6
|
|
|
5
|
-
|
|
6
|
-
[](https://www.npmjs.com/package/loki-mode)
|
|
7
|
-
[](https://github.com/asklokesh/loki-mode)
|
|
8
|
-
[](LICENSE)
|
|
9
|
-
[]()
|
|
10
|
-
[](https://www.autonomi.dev/)
|
|
11
|
-
[](https://hub.docker.com/r/asklokesh/loki-mode)
|
|
7
|
+
**PRD in, tested code out.**
|
|
12
8
|
|
|
13
|
-
|
|
9
|
+
[](https://www.npmjs.com/package/loki-mode)
|
|
10
|
+
[](https://www.npmjs.com/package/loki-mode)
|
|
11
|
+
[](https://github.com/asklokesh/loki-mode)
|
|
12
|
+
[](https://hub.docker.com/r/asklokesh/loki-mode)
|
|
13
|
+
[](LICENSE)
|
|
14
14
|
|
|
15
|
-
|
|
15
|
+
[Website](https://www.autonomi.dev/) | [Documentation](wiki/Home.md) | [Installation](docs/INSTALLATION.md) | [Changelog](CHANGELOG.md) | [Purple Lab Web UI](#purple-lab)
|
|
16
16
|
|
|
17
|
-
|
|
17
|
+
</div>
|
|
18
18
|
|
|
19
19
|
---
|
|
20
20
|
|
|
21
|
-
|
|
21
|
+
> **How it works:** You provide a PRD. Loki Mode classifies complexity, assembles an agent team from 41 specialized types across 8 swarms, and runs autonomous RARV cycles (Reason - Act - Reflect - Verify) with 9 quality gates. Code is not "done" until it passes automated verification. Output is a Git repo with source, tests, configs, and audit logs.
|
|
22
22
|
|
|
23
|
-
|
|
23
|
+
---
|
|
24
24
|
|
|
25
|
-
|
|
25
|
+
## Get Started in 30 Seconds
|
|
26
26
|
|
|
27
|
-
|
|
27
|
+
```bash
|
|
28
|
+
npm install -g loki-mode
|
|
29
|
+
loki doctor # verify environment
|
|
30
|
+
loki start ./prd.md # autonomous build starts
|
|
31
|
+
```
|
|
28
32
|
|
|
29
|
-
|
|
33
|
+
<details>
|
|
34
|
+
<summary><strong>Other install methods</strong></summary>
|
|
30
35
|
|
|
31
|
-
|
|
|
32
|
-
|
|
33
|
-
|
|
|
34
|
-
|
|
|
35
|
-
|
|
|
36
|
+
| Method | Command |
|
|
37
|
+
|--------|---------|
|
|
38
|
+
| **Homebrew** | `brew tap asklokesh/tap && brew install loki-mode` |
|
|
39
|
+
| **Docker** | `docker pull asklokesh/loki-mode` |
|
|
40
|
+
| **Inside Claude Code** | `claude --dangerously-skip-permissions` then type "Loki Mode" |
|
|
41
|
+
| **Git clone** | `git clone https://github.com/asklokesh/loki-mode.git` |
|
|
36
42
|
|
|
37
|
-
|
|
43
|
+
See the full [Installation Guide](docs/INSTALLATION.md).
|
|
38
44
|
|
|
39
|
-
|
|
40
|
-
|------|-----------|---------------------|
|
|
41
|
-
| **Code Generation** | Full-stack apps from PRDs | Complex domain logic may need human review |
|
|
42
|
-
| **Deployment** | Generates configs, Dockerfiles, CI/CD workflows | Does not deploy -- human provides cloud credentials and runs deploy |
|
|
43
|
-
| **Testing** | 9 automated quality gates, blind review | Test quality depends on AI-generated assertions |
|
|
44
|
-
| **Multi-Provider** | Claude (full), Codex/Gemini/Cline/Aider (sequential only) | Non-Claude providers lack parallel agents and Task tool |
|
|
45
|
-
| **Enterprise** | TLS, OIDC, RBAC, audit trail | Self-signed certs only; some features require env var activation |
|
|
46
|
-
| **Dashboard** | Real-time status, task queue, agents | Single-machine only; no multi-node clustering |
|
|
45
|
+
</details>
|
|
47
46
|
|
|
48
47
|
---
|
|
49
48
|
|
|
50
|
-
##
|
|
49
|
+
## What To Expect
|
|
51
50
|
|
|
52
|
-
|
|
51
|
+
| | Simple | Standard | Complex |
|
|
52
|
+
|---|---|---|---|
|
|
53
|
+
| **Examples** | Landing page, todo app, single API | CRUD + auth, REST API + React | Microservices, real-time, ML pipelines |
|
|
54
|
+
| **Duration** | 5-30 min | 30-90 min | 2+ hours |
|
|
55
|
+
| **Autonomy** | Completes independently | May need guidance on complex parts | Use as accelerator with human review |
|
|
53
56
|
|
|
54
|
-
|
|
57
|
+
---
|
|
55
58
|
|
|
56
|
-
|
|
57
|
-
npm install -g loki-mode
|
|
58
|
-
loki doctor # verify environment
|
|
59
|
-
loki start ./prd.md # uses Claude Code by default
|
|
60
|
-
```
|
|
59
|
+
## Architecture
|
|
61
60
|
|
|
62
|
-
|
|
61
|
+
<div align="center">
|
|
62
|
+
<img width="100%" alt="Loki Mode Architecture" src="https://github.com/user-attachments/assets/c9798120-9587-4847-8e8d-8f421f984dfc" />
|
|
63
|
+
</div>
|
|
63
64
|
|
|
64
|
-
|
|
65
|
-
|
|
66
|
-
|
|
67
|
-
```
|
|
65
|
+
<table>
|
|
66
|
+
<tr>
|
|
67
|
+
<td width="33%" valign="top">
|
|
68
68
|
|
|
69
|
-
|
|
69
|
+
### RARV Cycle
|
|
70
|
+
Every iteration: **Reason** (read state) - **Act** (execute, commit) - **Reflect** (update context) - **Verify** (run tests, check spec). Failures trigger self-correction.
|
|
70
71
|
|
|
71
|
-
|
|
72
|
+
[Core Workflow](references/core-workflow.md)
|
|
72
73
|
|
|
73
|
-
|
|
74
|
+
</td>
|
|
75
|
+
<td width="33%" valign="top">
|
|
74
76
|
|
|
75
|
-
|
|
77
|
+
### 41 Agent Types
|
|
78
|
+
8 swarms: engineering, operations, business, data, product, growth, review, orchestration. Auto-composed by PRD complexity.
|
|
76
79
|
|
|
77
|
-
|
|
80
|
+
[Agent Types](references/agent-types.md)
|
|
78
81
|
|
|
79
|
-
|
|
82
|
+
</td>
|
|
83
|
+
<td width="33%" valign="top">
|
|
80
84
|
|
|
81
|
-
|
|
85
|
+
### 9 Quality Gates
|
|
86
|
+
Blind review, anti-sycophancy, severity blocking, mock/mutation detection. Code does not ship until all gates pass.
|
|
82
87
|
|
|
83
|
-
|
|
88
|
+
[Quality Gates](skills/quality-gates.md)
|
|
84
89
|
|
|
85
|
-
|
|
86
|
-
|
|
87
|
-
|
|
90
|
+
</td>
|
|
91
|
+
</tr>
|
|
92
|
+
<tr>
|
|
93
|
+
<td width="33%" valign="top">
|
|
88
94
|
|
|
89
|
-
|
|
90
|
-
|
|
91
|
-
# Architecture decisions are injected as execution context
|
|
92
|
-
# Epics/stories are loaded into the task queue
|
|
93
|
-
```
|
|
95
|
+
### Memory System
|
|
96
|
+
3-tier architecture: episodic (interaction traces), semantic (generalized patterns), procedural (learned skills). Vector search optional.
|
|
94
97
|
|
|
95
|
-
|
|
98
|
+
[Memory Architecture](references/memory-system.md)
|
|
96
99
|
|
|
97
|
-
|
|
100
|
+
</td>
|
|
101
|
+
<td width="33%" valign="top">
|
|
98
102
|
|
|
99
|
-
|
|
103
|
+
### Dashboard
|
|
104
|
+
Real-time monitoring, agent status, task queue, WebSocket streaming. Auto-starts at `localhost:57374`.
|
|
100
105
|
|
|
101
|
-
|
|
106
|
+
[Dashboard Guide](docs/dashboard-guide.md)
|
|
102
107
|
|
|
103
|
-
|
|
108
|
+
</td>
|
|
109
|
+
<td width="33%" valign="top">
|
|
104
110
|
|
|
105
|
-
|
|
111
|
+
### Enterprise Layer
|
|
112
|
+
TLS, OIDC/SSO, RBAC, OTEL tracing, policy engine, audit trails. Activated via env vars.
|
|
113
|
+
|
|
114
|
+
[Enterprise Guide](docs/enterprise/architecture.md)
|
|
115
|
+
|
|
116
|
+
</td>
|
|
117
|
+
</tr>
|
|
118
|
+
</table>
|
|
106
119
|
|
|
107
120
|
---
|
|
108
121
|
|
|
109
|
-
##
|
|
122
|
+
## Purple Lab
|
|
123
|
+
|
|
124
|
+
The hosted development platform. A Replit-like web UI for visual PRD-to-code workflow with AI chat for iterative development.
|
|
110
125
|
|
|
111
|
-
|
|
126
|
+
```bash
|
|
127
|
+
loki web # launches at http://localhost:57375
|
|
128
|
+
```
|
|
112
129
|
|
|
130
|
+
<table>
|
|
131
|
+
<tr>
|
|
132
|
+
<td width="50%" valign="top">
|
|
113
133
|
|
|
114
|
-
|
|
134
|
+
**Platform Pages**
|
|
135
|
+
- Home -- PRD input with plan-before-build flow
|
|
136
|
+
- Projects -- browse, search, filter past builds
|
|
137
|
+
- Templates -- 20+ starter PRDs by category
|
|
138
|
+
- Settings -- provider selection, configuration
|
|
115
139
|
|
|
116
|
-
|
|
140
|
+
</td>
|
|
141
|
+
<td width="50%" valign="top">
|
|
117
142
|
|
|
118
|
-
**
|
|
143
|
+
**IDE Workspace**
|
|
144
|
+
- Monaco editor with tabs, Cmd+P quick open
|
|
145
|
+
- AI chat panel for iterative development
|
|
146
|
+
- Activity panel: build log, agents, quality gates
|
|
147
|
+
- Live preview with URL bar navigation
|
|
148
|
+
- Right-click context menu: Review, Test, Explain
|
|
119
149
|
|
|
120
|
-
|
|
121
|
-
|
|
122
|
-
|
|
123
|
-
- **Memory System** -- Episodic, semantic, procedural tiers with progressive disclosure. [Memory Architecture](references/memory-system.md)
|
|
124
|
-
- **Dashboard** -- Real-time monitoring, API v2, WebSocket at port 57374. [Dashboard Guide](docs/dashboard-guide.md)
|
|
125
|
-
- **Enterprise Layer** -- OTEL, policy engine, audit trails, RBAC, SSO (requires env var activation). [Enterprise Guide](docs/enterprise/architecture.md)
|
|
150
|
+
</td>
|
|
151
|
+
</tr>
|
|
152
|
+
</table>
|
|
126
153
|
|
|
127
154
|
---
|
|
128
155
|
|
|
129
|
-
##
|
|
130
|
-
|
|
131
|
-
|
|
|
132
|
-
|
|
133
|
-
| **
|
|
134
|
-
| **
|
|
135
|
-
| **
|
|
136
|
-
| **
|
|
137
|
-
| **
|
|
138
|
-
|
|
139
|
-
|
|
140
|
-
| **Deployment** | Helm, Docker Compose, Terraform configs (AWS/Azure/GCP) | [Deployment Guide](deploy/helm/README.md) |
|
|
141
|
-
| **Web App** | Replit-like UI with 10 React components, PRD input, agent dashboard, file browser, memory viewer | [Dashboard Guide](docs/dashboard-guide.md) |
|
|
142
|
-
| **Cost Estimation** | Pre-execution analysis with complexity scoring, token/cost projection | [Memory System](references/memory-system.md) |
|
|
143
|
-
| **Auto-Failover** | Cross-provider failover (Claude -> Codex -> Gemini) when rate limited | [Provider Guide](skills/providers.md) |
|
|
144
|
-
| **SDKs** | Python (`loki-mode-sdk`), TypeScript (`loki-mode-sdk`) | [SDK Guide](docs/enterprise/sdk-guide.md) |
|
|
145
|
-
|
|
146
|
-
### Multi-Provider Support
|
|
147
|
-
|
|
148
|
-
| Provider | Install | Autonomous Flag | Parallel Agents |
|
|
149
|
-
|----------|---------|-----------------|-----------------|
|
|
150
|
-
| Claude Code | `npm i -g @anthropic-ai/claude-code` | `--dangerously-skip-permissions` | Yes (10+) |
|
|
151
|
-
| Codex CLI | `npm i -g @openai/codex` | `--full-auto` | No (sequential) |
|
|
152
|
-
| Gemini CLI | `npm i -g @google/gemini-cli` | `--approval-mode=yolo` | No (sequential) |
|
|
153
|
-
| Cline CLI | `npm i -g @anthropic-ai/cline` | `--auto-approve` | No (sequential) |
|
|
154
|
-
| Aider | `pip install aider-chat` | `--yes-always` | No (sequential) |
|
|
155
|
-
|
|
156
|
-
Claude gets full features (subagents, parallelization, MCP, Task tool). All other providers run in sequential mode -- one agent at a time, no Task tool. See [Provider Guide](skills/providers.md) for the full comparison.
|
|
156
|
+
## Multi-Provider Support
|
|
157
|
+
|
|
158
|
+
| Provider | Autonomous Flag | Parallel Agents | Install |
|
|
159
|
+
|----------|:-:|:-:|---------|
|
|
160
|
+
| **Claude Code** | `--dangerously-skip-permissions` | Yes (10+) | `npm i -g @anthropic-ai/claude-code` |
|
|
161
|
+
| **Codex CLI** | `--full-auto` | Sequential | `npm i -g @openai/codex` |
|
|
162
|
+
| **Gemini CLI** | `--approval-mode=yolo` | Sequential | `npm i -g @google/gemini-cli` |
|
|
163
|
+
| **Cline CLI** | `--auto-approve` | Sequential | `npm i -g @anthropic-ai/cline` |
|
|
164
|
+
| **Aider** | `--yes-always` | Sequential | `pip install aider-chat` |
|
|
165
|
+
|
|
166
|
+
Claude gets full features (subagents, parallelization, MCP, Task tool). Other providers run sequentially. Auto-failover switches providers when rate-limited. See [Provider Guide](skills/providers.md).
|
|
157
167
|
|
|
158
168
|
---
|
|
159
169
|
|
|
160
|
-
## CLI
|
|
170
|
+
## CLI Reference
|
|
171
|
+
|
|
172
|
+
<details>
|
|
173
|
+
<summary><strong>All commands</strong></summary>
|
|
161
174
|
|
|
162
175
|
| Command | Description |
|
|
163
176
|
|---------|-------------|
|
|
@@ -166,67 +179,118 @@ Claude gets full features (subagents, parallelization, MCP, Task tool). All othe
|
|
|
166
179
|
| `loki pause` / `resume` | Pause/resume after current session |
|
|
167
180
|
| `loki status` | Show current status |
|
|
168
181
|
| `loki dashboard` | Open web dashboard |
|
|
182
|
+
| `loki web` | Launch Purple Lab web UI |
|
|
169
183
|
| `loki doctor` | Check environment and dependencies |
|
|
184
|
+
| `loki plan [PRD]` | Pre-execution analysis: complexity, cost, iterations |
|
|
185
|
+
| `loki review [--staged\|--diff]` | AI-powered code review with severity filtering |
|
|
186
|
+
| `loki test [--file\|--dir\|--changed]` | AI test generation (8 languages, 9 frameworks) |
|
|
187
|
+
| `loki onboard [path]` | Project analysis and CLAUDE.md generation |
|
|
170
188
|
| `loki import` | Import GitHub issues as tasks |
|
|
171
|
-
| `loki
|
|
172
|
-
| `loki
|
|
173
|
-
| `loki
|
|
174
|
-
| `loki
|
|
175
|
-
| `loki onboard [path]` | Instant project analysis and CLAUDE.md generation (12+ config types, 3 depth levels) |
|
|
176
|
-
| `loki ci` | CI/CD quality gate integration (GitHub Actions, GitLab CI, Jenkins, CircleCI) |
|
|
177
|
-
| `loki test [--file\|--dir\|--changed]` | AI-powered test generation (8 languages, 9 frameworks) |
|
|
178
|
-
| `loki failover [status\|--enable\|--chain]` | Cross-provider auto-failover when primary hits rate limits |
|
|
179
|
-
| `loki web` | Launch the web app (Replit-like UI for visual PRD-to-code workflow) |
|
|
189
|
+
| `loki ci` | CI/CD quality gate integration |
|
|
190
|
+
| `loki failover` | Cross-provider auto-failover management |
|
|
191
|
+
| `loki memory <cmd>` | Memory system: index, timeline, search, consolidate |
|
|
192
|
+
| `loki enterprise` | Enterprise feature management |
|
|
180
193
|
| `loki version` | Show version |
|
|
181
194
|
|
|
182
|
-
|
|
195
|
+
</details>
|
|
196
|
+
|
|
197
|
+
Run `loki --help` for all options. Full reference: [CLI Reference](wiki/CLI-Reference.md) | Config: [config.example.yaml](autonomy/config.example.yaml)
|
|
183
198
|
|
|
184
199
|
---
|
|
185
200
|
|
|
186
|
-
|
|
201
|
+
<details>
|
|
202
|
+
<summary><strong>BMAD Method Integration</strong></summary>
|
|
203
|
+
|
|
204
|
+
Loki Mode integrates with the [BMAD Method](https://github.com/bmad-code-org/BMAD-METHOD), a structured AI-driven agile methodology. If your project uses BMAD for requirements elicitation, Loki Mode can consume those artifacts directly:
|
|
205
|
+
|
|
206
|
+
```bash
|
|
207
|
+
loki start --bmad-project ./my-project
|
|
208
|
+
```
|
|
209
|
+
|
|
210
|
+
The adapter handles BMAD's frontmatter conventions, FR-format functional requirements, Given/When/Then acceptance criteria, and artifact chain validation. Non-BMAD projects are unaffected -- the integration is opt-in via `--bmad-project`.
|
|
187
211
|
|
|
188
|
-
|
|
212
|
+
See [BMAD Integration Validation](docs/architecture/bmad-integration-validation.md).
|
|
213
|
+
|
|
214
|
+
</details>
|
|
215
|
+
|
|
216
|
+
<details>
|
|
217
|
+
<summary><strong>Enterprise Features</strong></summary>
|
|
218
|
+
|
|
219
|
+
Enterprise features are included but require env var activation. Self-audit: 35/45 capabilities working, 0 broken, 1,314 tests passing.
|
|
189
220
|
|
|
190
221
|
```bash
|
|
191
222
|
export LOKI_TLS_ENABLED=true
|
|
192
223
|
export LOKI_OIDC_PROVIDER=google
|
|
193
224
|
export LOKI_AUDIT_ENABLED=true
|
|
194
|
-
|
|
195
|
-
loki enterprise status # check what's enabled
|
|
196
|
-
loki start ./prd.md # enterprise features activate via env vars
|
|
225
|
+
loki enterprise status
|
|
197
226
|
```
|
|
198
227
|
|
|
199
|
-
[Enterprise Architecture](docs/enterprise/architecture.md) | [Security](docs/enterprise/security.md) | [Authentication](docs/authentication.md) | [Authorization](docs/authorization.md) | [Metrics](docs/metrics.md) | [Audit Logging](docs/audit-logging.md)
|
|
228
|
+
[Enterprise Architecture](docs/enterprise/architecture.md) | [Security](docs/enterprise/security.md) | [Authentication](docs/authentication.md) | [Authorization](docs/authorization.md) | [Metrics](docs/metrics.md) | [Audit Logging](docs/audit-logging.md)
|
|
200
229
|
|
|
201
|
-
|
|
230
|
+
</details>
|
|
202
231
|
|
|
203
|
-
|
|
232
|
+
<details>
|
|
233
|
+
<summary><strong>Benchmarks</strong></summary>
|
|
204
234
|
|
|
205
|
-
|
|
235
|
+
Self-reported results from the included test harness. Verification scripts included for reproduction.
|
|
206
236
|
|
|
207
237
|
| Benchmark | Result | Notes |
|
|
208
238
|
|-----------|--------|-------|
|
|
209
|
-
| HumanEval | 162/164 (98.78%) | Max 3 retries
|
|
210
|
-
| SWE-bench | 299/300 patches
|
|
239
|
+
| HumanEval | 162/164 (98.78%) | Max 3 retries, RARV self-verification |
|
|
240
|
+
| SWE-bench | 299/300 patches | Patch generation -- evaluator not yet run |
|
|
241
|
+
|
|
242
|
+
See [benchmarks/](benchmarks/) for methodology.
|
|
243
|
+
|
|
244
|
+
</details>
|
|
245
|
+
|
|
246
|
+
<details>
|
|
247
|
+
<summary><strong>Presentation</strong></summary>
|
|
248
|
+
|
|
249
|
+
|
|
250
|
+
|
|
251
|
+
*9 slides: Problem, Solution, 41 Agents, RARV Cycle, Benchmarks, Multi-Provider, Full Lifecycle*
|
|
252
|
+
|
|
253
|
+
**[Download PPTX](docs/loki-mode-presentation.pptx)**
|
|
254
|
+
|
|
255
|
+
</details>
|
|
256
|
+
|
|
257
|
+
---
|
|
258
|
+
|
|
259
|
+
## Limitations
|
|
260
|
+
|
|
261
|
+
| Area | What Works | What Doesn't (Yet) |
|
|
262
|
+
|------|-----------|---------------------|
|
|
263
|
+
| **Code Gen** | Full-stack apps from PRDs | Complex domain logic may need human review |
|
|
264
|
+
| **Deploy** | Generates configs, Dockerfiles, CI/CD | Does not deploy -- human runs deploy commands |
|
|
265
|
+
| **Testing** | 9 automated quality gates | Test quality depends on AI assertions |
|
|
266
|
+
| **Providers** | 5 providers with auto-failover | Non-Claude providers lack parallel agents |
|
|
267
|
+
| **Dashboard** | Real-time single-machine monitoring | No multi-node clustering |
|
|
268
|
+
|
|
269
|
+
> **What "autonomous" means:** The system runs RARV cycles without prompting. It does NOT access your cloud accounts, payment systems, or external services unless you provide credentials. Human oversight is expected for deployment, API keys, and critical decisions.
|
|
211
270
|
|
|
212
271
|
---
|
|
213
272
|
|
|
214
273
|
## Research Foundation
|
|
215
274
|
|
|
216
|
-
|
|
217
|
-
|
|
218
|
-
|
|
275
|
+
<details>
|
|
276
|
+
<summary><strong>Papers and sources</strong></summary>
|
|
277
|
+
|
|
278
|
+
| Source | What We Use |
|
|
279
|
+
|--------|-------------|
|
|
280
|
+
| [Anthropic: Building Effective Agents](https://www.anthropic.com/research/building-effective-agents) | Evaluator-optimizer, parallelization |
|
|
219
281
|
| [Anthropic: Constitutional AI](https://www.anthropic.com/research/constitutional-ai-harmlessness-from-ai-feedback) | Self-critique against quality principles |
|
|
220
282
|
| [DeepMind: Scalable Oversight via Debate](https://deepmind.google/research/publications/34920/) | Debate-based verification in council review |
|
|
221
283
|
| [DeepMind: SIMA 2](https://deepmind.google/blog/sima-2-an-agent-that-plays-reasons-and-learns-with-you-in-virtual-3d-worlds/) | Self-improvement loop design |
|
|
222
|
-
| [OpenAI: Agents SDK](https://openai.github.io/openai-agents-python/) | Guardrails, tripwires, tracing
|
|
223
|
-
| [NVIDIA ToolOrchestra](https://github.com/NVlabs/ToolOrchestra) | Efficiency metrics, reward
|
|
224
|
-
| [CONSENSAGENT (ACL 2025)](https://aclanthology.org/2025.findings-acl.1141/) | Anti-sycophancy
|
|
284
|
+
| [OpenAI: Agents SDK](https://openai.github.io/openai-agents-python/) | Guardrails, tripwires, tracing |
|
|
285
|
+
| [NVIDIA ToolOrchestra](https://github.com/NVlabs/ToolOrchestra) | Efficiency metrics, reward signals |
|
|
286
|
+
| [CONSENSAGENT (ACL 2025)](https://aclanthology.org/2025.findings-acl.1141/) | Anti-sycophancy in blind review |
|
|
225
287
|
| [GoalAct](https://arxiv.org/abs/2504.16563) | Hierarchical planning for complex PRDs |
|
|
226
288
|
|
|
227
|
-
**Practitioner insights:** Boris Cherny
|
|
289
|
+
**Practitioner insights:** Boris Cherny, Simon Willison, [HN Community](https://news.ycombinator.com/item?id=44623207)
|
|
228
290
|
|
|
229
|
-
**[Full Acknowledgements](docs/ACKNOWLEDGEMENTS.md)** -- 50+
|
|
291
|
+
**[Full Acknowledgements](docs/ACKNOWLEDGEMENTS.md)** -- 50+ papers and resources
|
|
292
|
+
|
|
293
|
+
</details>
|
|
230
294
|
|
|
231
295
|
---
|
|
232
296
|
|
|
@@ -234,19 +298,20 @@ Results from the included test harness. Self-reported and not independently veri
|
|
|
234
298
|
|
|
235
299
|
```bash
|
|
236
300
|
git clone https://github.com/asklokesh/loki-mode.git && cd loki-mode
|
|
237
|
-
npm install && npm test # 683 tests
|
|
238
|
-
python3 -m pytest # 631 tests
|
|
239
|
-
bash tests/run-all-tests.sh # shell tests, ~2 min
|
|
301
|
+
npm install && npm test # 683 tests
|
|
302
|
+
python3 -m pytest # 631 tests
|
|
240
303
|
```
|
|
241
304
|
|
|
242
305
|
See [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines.
|
|
243
306
|
|
|
244
307
|
## License
|
|
245
308
|
|
|
246
|
-
Business Source License 1.1
|
|
247
|
-
|
|
248
|
-
Free for personal, internal, academic, and non-commercial use. Commercial use that competes with Loki Mode requires a separate license. Converts to Apache 2.0 on March 19, 2030. Contact founder@autonomi.dev for commercial licensing.
|
|
309
|
+
[Business Source License 1.1](LICENSE) -- Free for personal, internal, academic, and non-commercial use. Converts to Apache 2.0 on March 19, 2030. Contact founder@autonomi.dev for commercial licensing.
|
|
249
310
|
|
|
250
311
|
---
|
|
251
312
|
|
|
252
|
-
|
|
313
|
+
<div align="center">
|
|
314
|
+
|
|
315
|
+
**[Autonomi](https://www.autonomi.dev/)** | **[Documentation](wiki/Home.md)** | **[Changelog](CHANGELOG.md)** | **[Comparisons](references/competitive-analysis.md)**
|
|
316
|
+
|
|
317
|
+
</div>
|
package/SKILL.md
CHANGED
|
@@ -3,7 +3,7 @@ name: loki-mode
|
|
|
3
3
|
description: Multi-agent autonomous startup system. Triggers on "Loki Mode". Takes PRD to deployed product with minimal human intervention. Requires --dangerously-skip-permissions flag.
|
|
4
4
|
---
|
|
5
5
|
|
|
6
|
-
# Loki Mode v6.
|
|
6
|
+
# Loki Mode v6.44.0
|
|
7
7
|
|
|
8
8
|
**You are an autonomous agent. You make decisions. You do not ask questions. You do not stop.**
|
|
9
9
|
|
|
@@ -267,4 +267,4 @@ The following features are documented in skill modules but not yet fully automat
|
|
|
267
267
|
| Quality gates 3-reviewer system | Implemented (v5.35.0) | 5 specialist reviewers in `skills/quality-gates.md`; execution in run.sh |
|
|
268
268
|
| Benchmarks (HumanEval, SWE-bench) | Infrastructure only | Runner scripts and datasets exist in `benchmarks/`; no published results |
|
|
269
269
|
|
|
270
|
-
**v6.
|
|
270
|
+
**v6.44.0 | [Autonomi](https://www.autonomi.dev/) flagship product | ~260 lines core**
|
package/VERSION
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
6.
|
|
1
|
+
6.44.0
|
package/autonomy/loki
CHANGED
|
@@ -3175,9 +3175,6 @@ cmd_web_stop() {
|
|
|
3175
3175
|
curl -s --max-time 3 -X POST "http://127.0.0.1:${port}/api/session/stop" >/dev/null 2>&1 || true
|
|
3176
3176
|
sleep 1
|
|
3177
3177
|
|
|
3178
|
-
# Kill any orphaned loki-run processes (belt-and-suspenders)
|
|
3179
|
-
pkill -f "loki-run-" 2>/dev/null || true
|
|
3180
|
-
|
|
3181
3178
|
# Try PID file first
|
|
3182
3179
|
if [ -f "$PURPLE_LAB_PID_FILE" ]; then
|
|
3183
3180
|
local pid
|
|
@@ -3214,6 +3211,26 @@ cmd_web_stop() {
|
|
|
3214
3211
|
fi
|
|
3215
3212
|
|
|
3216
3213
|
rm -f "$PURPLE_LAB_PID_FILE" "${LOKI_DIR}/purple-lab/port" 2>/dev/null || true
|
|
3214
|
+
|
|
3215
|
+
# Kill orphaned loki-run processes and their entire process trees
|
|
3216
|
+
# This catches claude/codex/gemini child processes that survive parent death
|
|
3217
|
+
local orphan_pids
|
|
3218
|
+
orphan_pids=$(pgrep -f "loki-run-" 2>/dev/null || true)
|
|
3219
|
+
if [ -n "$orphan_pids" ]; then
|
|
3220
|
+
echo "Cleaning up orphaned build processes..."
|
|
3221
|
+
for opid in $orphan_pids; do
|
|
3222
|
+
# Kill the entire process tree rooted at this PID
|
|
3223
|
+
pkill -TERM -P "$opid" 2>/dev/null || true
|
|
3224
|
+
kill -TERM "$opid" 2>/dev/null || true
|
|
3225
|
+
done
|
|
3226
|
+
sleep 2
|
|
3227
|
+
# SIGKILL any survivors
|
|
3228
|
+
for opid in $orphan_pids; do
|
|
3229
|
+
pkill -9 -P "$opid" 2>/dev/null || true
|
|
3230
|
+
kill -9 "$opid" 2>/dev/null || true
|
|
3231
|
+
done
|
|
3232
|
+
echo "Orphaned build processes cleaned up."
|
|
3233
|
+
fi
|
|
3217
3234
|
}
|
|
3218
3235
|
|
|
3219
3236
|
cmd_web_status() {
|
package/dashboard/__init__.py
CHANGED
package/docs/INSTALLATION.md
CHANGED
package/package.json
CHANGED
|
@@ -1,4 +1,4 @@
|
|
|
1
|
-
import{c as m,r,j as e}from"./index-
|
|
1
|
+
import{c as m,r,j as e}from"./index-DKbfn_6W.js";import{C as g,a as p}from"./clock-KRe2srsu.js";/**
|
|
2
2
|
* @license lucide-react v0.577.0 - ISC
|
|
3
3
|
*
|
|
4
4
|
* This source code is licensed under the ISC license.
|
|
@@ -1 +1 @@
|
|
|
1
|
-
import{r as p,j as t}from"./index-
|
|
1
|
+
import{r as p,j as t}from"./index-DKbfn_6W.js";const m={primary:"bg-[#553DE9] text-white hover:bg-[#4432c4] shadow-button rounded-btn",secondary:"border border-[#553DE9] text-[#553DE9] hover:bg-[#E8E4FD] bg-transparent rounded-btn",ghost:"text-[#36342E] hover:bg-[#F8F4F0] rounded-btn",danger:"bg-[#C45B5B]/10 text-[#C45B5B] border border-[#C45B5B]/20 hover:bg-[#C45B5B]/20 rounded-btn"},b={sm:"px-3 py-1.5 text-xs",md:"px-4 py-2 text-sm",lg:"px-6 py-3 text-base"},u={sm:14,md:16,lg:18};function h({size:e}){return t.jsxs("svg",{className:"animate-spin",width:e,height:e,viewBox:"0 0 24 24",fill:"none",children:[t.jsx("circle",{className:"opacity-25",cx:"12",cy:"12",r:"10",stroke:"currentColor",strokeWidth:"4"}),t.jsx("path",{className:"opacity-75",fill:"currentColor",d:"M4 12a8 8 0 018-8V0C5.373 0 0 5.373 0 12h4z"})]})}const B=p.forwardRef(({variant:e="primary",size:o="md",icon:n,iconRight:i,loading:r=!1,disabled:a,className:c="",children:l,...x},d)=>{const s=u[o];return t.jsxs("button",{ref:d,disabled:a||r,className:["inline-flex items-center justify-center gap-2 font-medium transition-colors",m[e],b[o],(a||r)&&"opacity-60 cursor-not-allowed",c].filter(Boolean).join(" "),...x,children:[r?t.jsx(h,{size:s}):n?t.jsx(n,{size:s}):null,l,i&&!r&&t.jsx(i,{size:s})]})});B.displayName="Button";export{B};
|
|
@@ -1 +1 @@
|
|
|
1
|
-
import{j as s}from"./index-
|
|
1
|
+
import{j as s}from"./index-DKbfn_6W.js";const n={none:"p-0",sm:"p-3",md:"p-4",lg:"p-6"};function p({hover:e=!1,padding:d="md",className:t="",children:a,onClick:r}){return s.jsx("div",{role:r?"button":void 0,tabIndex:r?0:void 0,onClick:r,onKeyDown:r?o=>{(o.key==="Enter"||o.key===" ")&&(o.preventDefault(),r())}:void 0,className:["bg-white border border-[#ECEAE3] rounded-[5px] shadow-card",e&&"hover:shadow-card-hover transition-shadow duration-200",r&&"cursor-pointer",n[d],t].filter(Boolean).join(" "),children:a})}export{p as C};
|