feat: connect Pi events to independent source acceptance (no model calls)

README.md

@@ -72,7 +72,7 @@ Positive controls passed: 3/3
 
 This is a **scenario regression harness for the SACP engine**: it constructs claims and mock events directly, without running a coding agent or calling a model. It prints by default; `--output-dir .local/harness` explicitly writes reports. Continue through the [example guide](./examples/README.md) for deployment, refund, recovery and optional LangGraph demos.
 
-A separate real-agent evaluation layer is planned, with Pi as the first integration pilot. It is not implemented or benchmarked yet; see [ADR-0001](./docs/adr/0001-real-agent-evaluation-with-pi.md) for the decision, independent acceptance criteria and limits.
+The separate **Pi integration** now provides a no-model installation probe, JSON event replay and independent source-test acceptance. Try `python -m examples.pi_evaluation`: a scripted success claim passes for correct code and is blocked for incorrect code. This is not a real-agent benchmark; see the [integration guide](./docs/pi-integration.md) and [ADR-0001](./docs/adr/0001-real-agent-evaluation-with-pi.md).
 
 <a id="architecture"></a>
 

README.zh-CN.md

@@ -72,7 +72,7 @@ Positive controls passed: 3/3
 
 这里是 **SACP 引擎的场景回归测试（Harness）**：直接构造声明与模拟事件，不运行 coding agent，也不调用模型。默认只打印，显式传入 `--output-dir .local/harness` 才写报告。接着按[示例导航](./examples/README.zh-CN.md)体验部署、退款、恢复和可选 LangGraph demo。
 
-另计划新增真实 Agent 评测层，以 Pi 为首个接入试点。目前尚未实现，也没有对应 benchmark 结果；选择理由、独立验收标准与边界见 [ADR-0001](./docs/adr/0001-real-agent-evaluation-with-pi.md)。
+独立的 **Pi 接入层**已提供无模型安装探测、JSON 事件回放和交付源码的独立测试验收。运行 `python -m examples.pi_evaluation`：同一条合成成功声明，正确源码通过、错误源码被阻止。尚无真实 Agent benchmark；用法与边界见[接入指南](./docs/pi-integration.md)和 [ADR-0001](./docs/adr/0001-real-agent-evaluation-with-pi.md)。
 
 <a id="框架"></a>
 

ROADMAP.md

@@ -12,7 +12,7 @@ CI runs a cross-platform core matrix, explicit LangGraph integration, strict pro
 
 ## Next: real-agent evaluation pilot
 
-[ADR-0001](./docs/adr/0001-real-agent-evaluation-with-pi.md) retains the existing deterministic engine scenarios and selects Pi as the first real-agent pilot. Implement a Python outer runner using Pi's JSON subprocess interface, isolate and pin its environment, then collect actual coding-task trajectories and independently verify the delivered patches. The adapter, provider/model, budget and real results are not yet available.
+[ADR-0001](./docs/adr/0001-real-agent-evaluation-with-pi.md) retains deterministic engine scenarios and selects Pi as the first real-agent pilot. The [initial integration](./docs/pi-integration.md) provides a no-model RPC probe, JSON subprocess capture API with default-denied model execution, event replay and independent acceptance of explicit Python fixture files. Next enforce isolation, pin the complete environment, select a provider/model and budget, then collect actual coding-task trajectories and verify delivered patches. No real-agent task result is available yet.
 
 Start with a small fixed task set and compare SACP with a predefined test-based baseline on the same records. Report missed false completions, incorrect blocks, indeterminate outcomes and overhead; separate injected faults from naturally occurring failures. Keep model calls out of required PR CI. Acceptance stages and stop conditions are defined in the ADR.
 

ROADMAP.zh-CN.md

@@ -12,7 +12,7 @@ CI 已运行跨平台 core 矩阵、显式安装依赖的 LangGraph 集成、严
 
 ## 下一步：真实 Agent 评测试点
 
-[ADR-0001](./docs/adr/0001-real-agent-evaluation-with-pi.md)决定保留现有确定性引擎场景测试，以 Pi 为首个真实 Agent 试点。计划使用 Python 外层 runner 与 Pi JSON 子进程接口，固定并隔离环境，采集实际代码任务轨迹并独立验收交付补丁。目前尚无 adapter、选定的 provider/model、运行预算或真实评测结果。
+[ADR-0001](./docs/adr/0001-real-agent-evaluation-with-pi.md)决定保留现有确定性引擎场景测试，以 Pi 为首个真实 Agent 试点。[初版接入](./docs/pi-integration.md)已有无模型 RPC 探测、默认禁止模型执行的 JSON 子进程采集 API、事件回放及明确列出的 Python fixture 文件独立验收。下一步落实强制隔离、完整环境锁定、provider/model 与预算，再采集实际代码任务轨迹并验收补丁。目前没有真实 Agent 任务结果。
 
 先用小规模固定任务集，在同一批记录上比较 SACP 与预先定义的测试基线，报告漏检、误拦、不可判定和额外开销；注入故障与自然失败分别统计。模型调用不进入 required PR CI。分阶段验收与停止条件见 ADR。
 

STATUS.md

@@ -16,12 +16,13 @@ Start with [README](./README.md) / [中文](./README.zh-CN.md), then [architectu
 | Adapters | REST-shaped GitHub payload mapping, trusted reviewer allowlist, CLI, real LangGraph wiring/checkpointers | Payload source authentication remains an integration responsibility |
 | HTTP experiments | Real localhost provider transport, simulated health/rollback | No public SACP API or production provider |
 | Scenario regression harness | Eight synthetic breakers and three positive controls for the SACP engine | Does not run a real agent/model; no measured population accuracy |
+| Pi evaluation integration | No-model RPC probe, default-denied JSON capture API, replay and independent Python fixture acceptance | No model task run, authenticated collector or hostile-code sandbox; see [guide](./docs/pi-integration.md) |
 | Governance | Read-only PR CI, trusted independent attestation, formal receipt, manual squash merge, main revalidation | Semantic review required; real fork E2E still outstanding |
 | Corpus and Doctor | 34 translated receipts, YAML examples, local text-audit CLI | Historical/synthetic inputs; Doctor uses heuristics and does not rerun tasks |
 
 ## Verification baseline
 
-Code baseline [1b76ff3](https://github.com/aDragon0707/sacp/commit/1b76ff386d6091535aae94acbb7517c9170bd5d8): 100 core tests, 7 LangGraph tests, 49 strict protocol documents, 8 breakers and 3 positive controls. Core CI covers Ubuntu Python 3.11/3.12 and Windows Python 3.12; LangGraph runs on Ubuntu Python 3.12. Zero failures/errors/unexpected skips are required.
+Historical code baseline [1b76ff3](https://github.com/aDragon0707/sacp/commit/1b76ff386d6091535aae94acbb7517c9170bd5d8): 100 core tests, 7 LangGraph tests, 49 strict protocol documents, 8 breakers and 3 positive controls. The Pi integration increases the core inventory to 136, including 35 no-model Pi tests and one classification test. Local execution on 2026-10-15 passed all four suites with no failures/errors/skips; SHA-bound CI artifacts remain the evidence for each merged revision. Core CI covers Ubuntu Python 3.11/3.12 and Windows Python 3.12; LangGraph runs on Ubuntu Python 3.12.
 
 See the [dated verification record](./docs/verification-2026-10-15.md) for exact runs and smoke coverage. A successful receipt-generation job can write a BLOCK or FAIL receipt; inspect content and current SHA, not just a green job.
 
@@ -33,6 +34,6 @@ The protocol itself does not require a database or HTTP transport. The Python re
 
 ## Next work
 
-The accepted [real-agent evaluation decision](./docs/adr/0001-real-agent-evaluation-with-pi.md) preserves deterministic scenarios and selects Pi as the first coding-agent pilot, using an outer runner and independent task acceptance. This is planned work: no Pi adapter, selected provider/model or real-agent benchmark result exists yet.
+The accepted [real-agent evaluation decision](./docs/adr/0001-real-agent-evaluation-with-pi.md) now has an [initial no-model integration](./docs/pi-integration.md). Actual Pi 1.0.0 `get_state` passed without prompts; deterministic replay and independent source acceptance exercise the evidence path. A selected provider/model, enforced execution isolation, budget and real-agent benchmark results remain future work.
 
 Finish controlled real fork end-to-end acceptance, strengthen isolation against arbitrary malicious Python tests, authenticate external collectors where used, and gather independently reviewed real-world false-completion and false-block cases. Keep schema/transport expansion behind concrete dirty cases and compatibility review. See [ROADMAP](./ROADMAP.md).

docs/adr/0001-real-agent-evaluation-with-pi.md

@@ -7,7 +7,7 @@ date: 2026-10-15
 
 当前 `sacp_harness/` 直接构造 Claim、mock provider 事件和授权，验证 SACP 引擎及协议文档；没有驱动真实 Agent 或调用模型。我们决定保留它作为确定性场景回归测试，另增真实 Agent 端到端评测层，并选择 Pi coding agent 作为首个有界接入试点，以回答 SACP 相比已有测试检查增加了什么判别价值。
 
-本 ADR 接受的是架构方向和首个试点对象。Pi adapter、执行环境、任务集和评测结果均尚未实现；源码可接入性已核对，安装兼容性、隔离和实际采集须经后续 spike 验收。
+本 ADR 最初接受的是架构方向和首个试点对象；当时 Pi adapter、执行环境、任务集和评测结果均尚未实现。后续进展独立记录于文末，保留以下原始决策与源码证据边界。
 
 ## 决策与术语
 
@@ -35,6 +35,12 @@ date: 2026-10-15
 
 选择 Pi 的理由是可观察和易接入，不是已有证据表明它的任务成功率优于其他 Agent。
 
+### 实施进展：2026-10-15，无模型接入阶段
+
+新增 `sacp_eval/`：JSON 事件解析、默认拒绝模型执行的子进程采集 API、独立源码 fixture 验收与 SACP 证据绑定。本机已安装的 Pi 1.0.0 通过 `--version` / RPC `get_state` 探测，未发送 prompt、未继承模型凭据；这不代表原始审查的 1.2.0 版本已安装或完整运行时已锁定。RPC 此处只用于状态探测，任务执行接缝仍为 JSON subprocess。
+
+合成轨迹与真实测试子进程已能区分正确/错误交付，并阻止旧源码证据复用；没有真实模型轨迹或效果指标。强制隔离、provider/model、预算和完整环境锁定尚未落实。用户明确要求先完成接入、暂不调用模型，因此完整真实 Agent pilot 仍待后续授权和环境准备。用法、接口与信任边界见[Pi 接入指南](../pi-integration.md)。
+
 ## 边界与执行流程
 
 ```mermaid

docs/architecture.md

@@ -75,7 +75,7 @@ The engine runs during a workflow to derive state and gate actions. The existing
 
 Additional tests cover cross-receipt identity, callback permutations, approval revocation, SQLite restart/concurrency, and accepted-but-not-terminal reconciliation. These tests do not establish a real-world error rate. The harness's trace column is explanatory text, not a measurement of tracing systems.
 
-[ADR-0001](./adr/0001-real-agent-evaluation-with-pi.md) accepts a separate real-agent evaluation layer, starting with Pi and a fixed model configuration. This layer is planned, not implemented. Pi is the agent harness under observation; the outer evaluation harness owns collection and independent task acceptance. LangGraph experiments remain workflow integration checks, not proof of autonomous-agent performance. The new layer will compare SACP with a test-based baseline on the same real trajectories while preserving deterministic PR CI. See [evaluation vocabulary](../GLOSSARY.md).
+[ADR-0001](./adr/0001-real-agent-evaluation-with-pi.md) accepts a separate real-agent evaluation layer. Its initial `sacp_eval/` implementation includes a no-model Pi probe, default-denied subprocess capture seam, JSON lifecycle parser and independent acceptance of explicit Python files. The trusted outer caller binds trace and delivery digests before recording a host test observation and selected claim. Pi is the agent harness under observation; the outer evaluator owns collection and acceptance. Current demos replay scripted events; no real model-driven task has run. Real trajectories, isolated execution and baseline comparison remain future work. LangGraph experiments remain workflow wiring checks. See the [integration guide](./pi-integration.md) and [evaluation vocabulary](../GLOSSARY.md).
 
 ## Why another layer?
 

docs/development.md

@@ -22,7 +22,7 @@ python -m unittest tests.test_harness -v
 git diff --check
 ```
 
-The code baseline linked from [README](../README.md) has 100 core tests, 7 integration tests, 49 protocol documents, 8 breakers and 3 positive controls. Counts are a versioned inventory, not a fixed success criterion. New tests should increase coverage; a missing dependency or skip is not a complete pass.
+The historical code baseline linked from [README](../README.md) has 100 core tests, 7 integration tests, 49 protocol documents, 8 breakers and 3 positive controls. The Pi integration increases core discovery to 136 tests. Counts are a versioned inventory, not a fixed success criterion. New tests should increase coverage; a missing dependency or skip is not a complete pass.
 
 The core suite discovers `tests/test_*.py`, excluding the optional LangGraph and harness modules. `--suite harness` executes the breaker/positive report directly; `tests.test_harness` separately checks report output and deterministic artifact handling. Protocol validates root YAML examples and translated corpus receipts strictly.
 
@@ -62,7 +62,7 @@ Formal CI receipts require an independent attestation from default-branch code.
 
 `python -m sacp_harness` remains a deterministic regression command for the SACP engine and protocol validator. It does not execute Pi, another coding agent or any model. LangGraph suites test graph/checkpointer wiring; they do not establish model-driven task performance.
 
-[ADR-0001](./adr/0001-real-agent-evaluation-with-pi.md) defines the planned Pi pilot: a separate Python outer runner, JSON subprocess capture, isolated/pinned runtime and independent acceptance on delivered code. No install command, agent run or new evaluation suite is implemented by that decision. Pi provider/model, runtime pin, budget and environment must be resolved before real execution. Ordinary core/protocol/harness jobs must retain their existing deterministic interface and must not silently invoke a model.
+[ADR-0001](./adr/0001-real-agent-evaluation-with-pi.md) defines the Pi pilot. The [initial integration](./pi-integration.md) now supports `python -m examples.pi_evaluation` and `python -m unittest tests.test_pi_evaluation -v`, without Pi installation, model calls or credentials. Core discovery includes these tests. Only the opt-in `probe` command uses an installed Pi, for version and RPC state. Real model runs still require a selected provider/model, budget, complete runtime pin and enforced isolation; the capture library defaults to refusing model execution. Local acceptance is for trusted fixtures only. Ordinary required CI retains its deterministic interface.
 
 ## Public/private boundary
 

docs/maintainer-workflow.md

@@ -31,6 +31,7 @@ The policy checker records zero or more classes in CI evidence and the formal re
 |---|---|---|
 | `protocol` | protocol documents and YAML packets | semantics, compatibility, dirty case, example, changelog |
 | `engine` | `sacp_verify/`, `sacp_harness/` | conservative state projection, append-only history, failure tests |
+| `adapter` | `sacp_eval/`, Pi evaluation examples | event contract, delivery binding, independent acceptance and explicit execution boundaries |
 | `adapter` | GitHub, LangGraph, and CLI boundaries | external facts must not be promoted to stronger states |
 | `automation` | workflows and CI scripts | permissions, action pins, self-modification, evidence binding |
 | `corpus` | `sample-corpus/` | privacy, contamination, reproducibility |