Skip to content

Commit dae2096

Browse files
Prepare public onboarding for v0.2.0
1 parent 5856a28 commit dae2096

12 files changed

Lines changed: 477 additions & 29 deletions

File tree

Lines changed: 43 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,43 @@
1+
name: RFC
2+
description: Propose a protocol, evidence, adapter, or governance change for review.
3+
title: "[RFC] "
4+
body:
5+
- type: markdown
6+
attributes:
7+
value: |
8+
Use RFCs for changes that affect protocol semantics, evidence contracts, runtime adapter expectations, schemas, or public governance. Use issues for scoped bugs.
9+
- type: textarea
10+
id: problem
11+
attributes:
12+
label: Problem
13+
description: What real agent workflow failure, adapter gap, or evidence ambiguity motivates this RFC?
14+
validations:
15+
required: true
16+
- type: textarea
17+
id: proposal
18+
attributes:
19+
label: Proposal
20+
description: What should change in SPEC, docs, schemas, CLI behavior, examples, or community process?
21+
validations:
22+
required: true
23+
- type: textarea
24+
id: evidence
25+
attributes:
26+
label: Evidence impact
27+
description: Which receipts, expected evidence, audit gates, approval gates, recommendation resolution, or adapter fields are affected?
28+
validations:
29+
required: true
30+
- type: textarea
31+
id: compatibility
32+
attributes:
33+
label: Compatibility
34+
description: Is this backwards-compatible? Does it affect Full Mode, Remote Mode, Manual Mode, or existing examples?
35+
validations:
36+
required: true
37+
- type: textarea
38+
id: smallest
39+
attributes:
40+
label: Smallest testable artifact
41+
description: What is the smallest doc patch, schema fixture, task folder, or CLI test that would prove the proposal?
42+
validations:
43+
required: false

CHANGELOG.md

Lines changed: 16 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -2,6 +2,22 @@
22

33
## Unreleased
44

5+
- No unreleased changes yet.
6+
7+
## 0.2.0
8+
9+
- Promotes the `0.2.0-draft` work into the first stable evaluation release.
10+
- Adds a root `README.zh-CN.md` Chinese entry point.
11+
- Adds a failure gallery that shows concrete false-done failure shapes and the
12+
VALP evidence that catches them.
13+
- Adds a runtime adapter checklist for adapter authors evaluating Full Mode,
14+
Remote Mode, or Manual Mode compatibility.
15+
- Adds a GitHub Discussions RFC template for protocol, evidence, schema,
16+
adapter, and governance proposals.
17+
- Documents that dispatch payloads are a coordinator/leader responsibility:
18+
concise worker prompts should cite task-local refs instead of pasting long
19+
context or full skill-router output.
20+
521
- Adds a project status and evidence matrix that separates current repository
622
proof from unproven live-runtime, adapter, case-study, and production claims.
723
- Clarifies that HERDR is the public reference runtime and not a VALP protocol

README.md

Lines changed: 17 additions & 8 deletions
Original file line numberDiff line numberDiff line change
@@ -6,8 +6,11 @@ Open protocol for visible, evidence-backed, multi-agent automation.
66

77
![VALP social preview: Agent says done, VALP asks for proof](docs/assets/social-preview.png)
88

9-
中文注解入口: [docs/zh-CN/README.md](docs/zh-CN/README.md). The Chinese pages
10-
are explanatory notes; `SPEC.md` and `schemas/` remain the normative protocol
9+
Languages: [English](README.md) | [中文](README.zh-CN.md)
10+
11+
中文注解入口: [README.zh-CN.md](README.zh-CN.md) and
12+
[docs/zh-CN/README.md](docs/zh-CN/README.md). The Chinese pages are
13+
explanatory notes; `SPEC.md` and `schemas/` remain the normative protocol
1114
source.
1215

1316
The protocol is designed for terminal-based AI coding agents, review agents,
@@ -16,9 +19,9 @@ single project, operating system, terminal emulator, or model provider.
1619

1720
## Current Status
1821

19-
VALP is a `0.2.0-draft` open protocol plus an MIT-licensed reference CLI. It is
20-
not a mature hosted platform and should not be described as production-ready by
21-
itself.
22+
VALP is a `0.2.0` open protocol release plus an MIT-licensed reference CLI. It
23+
is not a mature hosted platform and should not be described as production-ready
24+
by itself.
2225

2326
What this repository proves today:
2427

@@ -71,7 +74,8 @@ Choose the path that matches why you are here:
7174
| Try automated multi-agent work | Install HERDR, the current reference runtime | Yes |
7275
| Enable automatic visible task intake | Read [docs/auto-visible-mode.md](docs/auto-visible-mode.md) | For dispatch, yes |
7376
| Inspect a headless runtime shape | Audit `examples/headless-queue-task/` | No |
74-
| Implement a new runtime | Read [docs/runtime-adapters.md](docs/runtime-adapters.md) | Depends on your adapter |
77+
| See what failures VALP catches | Read [docs/failure-gallery.md](docs/failure-gallery.md) | No |
78+
| Implement a new runtime | Read [docs/adapter-checklist.md](docs/adapter-checklist.md) and [docs/runtime-adapters.md](docs/runtime-adapters.md) | Depends on your adapter |
7579

7680
## Community
7781

@@ -81,6 +85,7 @@ runtime evidence. The project is looking for:
8185
- runtime adapter feedback from queues, hosted agent systems, terminal
8286
controllers, and manual review workflows;
8387
- small audited examples that show where visible receipts help or fail;
88+
- RFCs for protocol, evidence, schema, adapter, or governance changes;
8489
- documentation improvements that make first install and first audit easier;
8590
- skeptical critiques of whether the protocol is useful or just ceremony.
8691

@@ -95,6 +100,7 @@ The best early feedback is not a generic "looks good". It is one of:
95100
- a false-done case where an agent or runtime claimed completion without proof;
96101
- a minimal audit run that failed or felt too ceremonial;
97102
- a runtime-adapter sketch that preserves receipt and evidence semantics.
103+
- an RFC that proposes the smallest evidence-changing protocol improvement.
98104

99105
No-runtime first look:
100106

@@ -424,6 +430,7 @@ receipt ledger
424430
```text
425431
Visible-Agent-Loop-Protocol/
426432
README.md
433+
README.zh-CN.md
427434
SPEC.md
428435
INSTALL.md
429436
ROADMAP.md
@@ -447,6 +454,8 @@ Visible-Agent-Loop-Protocol/
447454
quickstart.md
448455
faq.md
449456
comparison.md
457+
failure-gallery.md
458+
adapter-checklist.md
450459
runtime-adapters.md
451460
community.md
452461
project-status.md
@@ -520,5 +529,5 @@ Visible-Agent-Loop-Protocol/
520529
## Status
521530

522531
Open protocol draft with reference CLI version `0.2.0`. The protocol draft is
523-
`0.2.0-draft`; HERDR remains the current reference runtime, not a protocol
524-
requirement.
532+
released as `v0.2.0`; HERDR remains the current reference runtime, not a
533+
protocol requirement.

README.zh-CN.md

Lines changed: 71 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,71 @@
1+
# Visible Agent Loop Protocol 中文入口
2+
3+
本页是中文入口,不是协议规范原文。若本页与 `SPEC.md``schemas/`
4+
`valp audit` 行为冲突,以英文规范和机器可验证规则为准。
5+
6+
VALP 是一个面向多 Agent 工作流的可见证据协议。它解决的问题很窄:
7+
8+
```text
9+
Agent 或 Runtime 说 done,用户凭什么相信?
10+
```
11+
12+
VALP 要求任务过程留下可审计的证据:
13+
14+
- 谁被分配了任务;
15+
- dispatch 有没有真的提交;
16+
- 预期证据是什么;
17+
- 证据文件是否存在;
18+
- 验证、审查、审批是否通过;
19+
- 最终结论引用了哪些证据。
20+
21+
所以 VALP 更像一个多 Agent 工作流的验收单,而不是聊天提示词集合。
22+
23+
## 五分钟体验
24+
25+
不需要先安装 Runtime:
26+
27+
```bash
28+
git clone https://github.com/wcqxgjy6d8-pixel/Visible-Agent-Loop-Protocol.git
29+
cd Visible-Agent-Loop-Protocol
30+
python -m pip install -r requirements-dev.txt
31+
bin/valp audit examples/minimal-task
32+
```
33+
34+
通过时你会看到:
35+
36+
```text
37+
VALP audit: PASS
38+
Summary: pass=13 warn=0 fail=0 skip=7
39+
```
40+
41+
再看 [Minimal audit demo](docs/minimal-audit-demo.md),它会展示:
42+
删掉 expected evidence 后,即使 receipt 说结果已提交,`valp audit` 也会失败。
43+
44+
## 它不是什么
45+
46+
VALP 不是:
47+
48+
- 一个托管平台;
49+
- 一个模型集成方法;
50+
- 一个固定绑定 HERDR 的私有工作流;
51+
- 一个能自动证明所有 Agent 可靠的魔法层;
52+
- 一个替代测试、代码审查、审批流程的工具。
53+
54+
HERDR 只是当前参考 Runtime。其他 Runtime 只要能导出同等的 receipts、
55+
evidence、state mapping 和 audit 数据,也可以实现 VALP。
56+
57+
## 推荐阅读
58+
59+
1. [英文 README](README.md)
60+
2. [协议规范 SPEC.md](SPEC.md)
61+
3. [中文注解](docs/zh-CN/README.md)
62+
4. [失败案例图鉴](docs/failure-gallery.md)
63+
5. [Runtime adapter checklist](docs/adapter-checklist.md)
64+
6. [社区参与说明](docs/community.md)
65+
66+
如果你要实现 Runtime adapter,优先读:
67+
68+
- [Runtime adapters](docs/runtime-adapters.md)
69+
- [Adapter checklist](docs/adapter-checklist.md)
70+
- [Dispatch receipts](docs/dispatch-receipts.md)
71+
- [Correction cycle evidence](docs/correction-cycle.md)

docs/adapter-checklist.md

Lines changed: 141 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,141 @@
1+
# Runtime Adapter Checklist
2+
3+
Use this checklist when evaluating whether a runtime can implement VALP.
4+
5+
The question is not "can this runtime run an agent?" The question is whether it
6+
can export enough evidence for another person, agent, or CI job to audit the
7+
work.
8+
9+
## 1. Classify The Adapter
10+
11+
Choose the closest shape:
12+
13+
| Adapter class | Typical runtime | Mode |
14+
|---|---|---|
15+
| pane controller | terminal panes or browser-controlled panes | Full Mode if proof is exported |
16+
| daemon queue | local job queue or runner daemon | Full Mode if state and evidence are exported |
17+
| hosted/local platform | web board plus agent workers | Full Mode if audit data is accessible |
18+
| remote SSH | runtime state lives on another host | Remote Mode |
19+
| manual | human copies prompts and results | Manual Mode |
20+
21+
## 2. Required Full Mode Evidence
22+
23+
A Full Mode adapter must export:
24+
25+
- agent list;
26+
- agent metadata and status;
27+
- provider matrix;
28+
- context policy;
29+
- runtime preflight;
30+
- dispatch submission proof;
31+
- runtime task state mapping;
32+
- expected evidence refs;
33+
- receipt ledger;
34+
- failure reason when blocked or failed;
35+
- approval gate status.
36+
37+
If any of these are missing, the adapter may still be useful, but it should not
38+
claim Full Mode yet.
39+
40+
## 3. Dispatch Proof
41+
42+
Check each dispatch step:
43+
44+
- `dispatch_written`: the task-local dispatch file or equivalent exists.
45+
- `dispatch_inserted`: text was inserted into an input surface, if applicable.
46+
- `dispatch_submitted`: the runtime proves the message/job was submitted.
47+
- `dispatch_completed`: expected evidence exists after execution.
48+
- `dispatch_blocked`: delivery or completion could not be proven.
49+
50+
Text inserted into a pane is not enough. Queue accepted is not enough. Hosted
51+
job completed is not enough. The adapter must connect runtime state to expected
52+
evidence.
53+
54+
## 4. Expected Evidence
55+
56+
For each selected agent, record:
57+
58+
- exact expected evidence paths;
59+
- who owns each path;
60+
- whether the evidence exists;
61+
- whether evidence is valid, superseded, rejected, blocked, or invalid;
62+
- whether build/test/runtime claims cite concrete logs or screenshots.
63+
64+
## 5. Preflight
65+
66+
Pane adapters should record:
67+
68+
- pane id;
69+
- foreground cwd, when available;
70+
- terminal size, when available;
71+
- minimum terminal size expected by the agent;
72+
- CLI availability and version probe, when available;
73+
- known TUI/display caveats.
74+
75+
Queue, hosted, or remote adapters should record equivalent job/session facts:
76+
77+
- queue id or run id;
78+
- worker id;
79+
- session status;
80+
- output refs;
81+
- retry state or failure reason;
82+
- expected refs.
83+
84+
Do not fake pane fields for non-pane runtimes.
85+
86+
## 6. Approval And Risk
87+
88+
The adapter must stop or require approval for:
89+
90+
- destructive changes;
91+
- release, publish, upload, submit, or deploy;
92+
- auth, secrets, signing, entitlements, privacy metadata;
93+
- migrations;
94+
- memory, agent config, or MCP config changes;
95+
- private-data export.
96+
97+
Approval evidence must be task-local and visible. A permissive runtime default
98+
does not override VALP approval gates.
99+
100+
## 7. Recommendation Resolution
101+
102+
If selected workers return recommendations:
103+
104+
- collect them in task-local evidence;
105+
- record whether each meaningful recommendation was adopted, merged, bounded,
106+
converted to follow-up, rejected, or escalated;
107+
- avoid unbounded redispatch loops;
108+
- require user approval when a recommendation expands into high-risk work.
109+
110+
## 8. Minimal Smoke Test
111+
112+
Before claiming compatibility, run a small task through this sequence:
113+
114+
```text
115+
publish task
116+
scan capabilities
117+
route selected agents
118+
write concise dispatch
119+
submit through runtime
120+
record dispatch_submitted proof
121+
wait for worker output
122+
write expected evidence
123+
record dispatch_completed
124+
write final synthesis
125+
run valp audit
126+
```
127+
128+
The task is compatible only when the audit passes for the right reason.
129+
130+
## 9. What To Post For Feedback
131+
132+
In a runtime-adapter discussion, include:
133+
134+
- adapter class;
135+
- runtime or platform name;
136+
- which Full Mode fields are exported today;
137+
- which fields are missing;
138+
- one sample receipt ledger;
139+
- one expected evidence map;
140+
- known failure modes;
141+
- whether the adapter should start as Full Mode, Remote Mode, or Manual Mode.

docs/community.md

Lines changed: 10 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -26,6 +26,8 @@ scripts/verify-examples.sh
2626
Use GitHub Discussions for open-ended feedback:
2727

2828
- whether VALP is useful reliability engineering or unnecessary ceremony;
29+
- RFCs for protocol semantics, schema, evidence, adapter, or governance
30+
changes;
2931
- runtime adapter ideas before they are ready for an issue;
3032
- task folder or audit experiments that need interpretation;
3133
- workflow stories where a runtime said "completed" before evidence existed.
@@ -45,6 +47,7 @@ The most useful community feedback is concrete:
4547
- a runtime adapter that can or cannot export VALP receipts and evidence;
4648
- an example task folder that passes or fails audit for a clear reason;
4749
- schema or documentation wording that produces the wrong mental model;
50+
- an RFC with a small evidence-changing proposal and a testable artifact;
4851
- first-install friction found by running `valp doctor`, preflight, or the
4952
quickstart.
5053

@@ -54,14 +57,17 @@ Good first contributions usually fit one of these shapes:
5457

5558
- improve quickstart or install wording;
5659
- add a small Manual Mode example;
57-
- add a runtime adapter capability checklist;
60+
- add or refine a runtime adapter capability checklist;
61+
- add a failure gallery entry with receipts, expected evidence, and audit
62+
behavior;
5863
- improve `valp doctor` or audit messages;
5964
- tighten schema/example consistency;
6065
- document a comparison against a real agent workflow.
6166

62-
For broad ideas, use GitHub Discussions. For concrete bugs or scoped changes,
63-
open an issue. For code or doc changes, open a pull request and include the
64-
verification command you ran.
67+
For broad ideas, use GitHub Discussions. Use the RFC template when the proposal
68+
changes protocol semantics, evidence contracts, schemas, adapter requirements,
69+
or governance. For concrete bugs or scoped changes, open an issue. For code or
70+
doc changes, open a pull request and include the verification command you ran.
6571

6672
## Ground Rules
6773

0 commit comments

Comments
 (0)