@piagent/platform 1.6.0 → 1.6.1

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.
Files changed (213) hide show
  1. package/CHANGELOG.md +49 -0
  2. package/README.md +17 -14
  3. package/README.vi.md +26 -1
  4. package/benchmarks/production-v1/spend-control.v1.json +27 -3
  5. package/benchmarks/production-v1/suite.json +5 -1
  6. package/catalog/capabilities.json +16 -16
  7. package/docs/authority-manifest.md +1 -1
  8. package/docs/auto-delegation-policy.md +38 -73
  9. package/docs/command-reference-vietnamese.md +23 -31
  10. package/docs/context-engine.md +5 -3
  11. package/docs/context-window-policy.md +37 -1
  12. package/docs/distribution-standard.md +7 -7
  13. package/docs/en/architecture.md +26 -13
  14. package/docs/languages.json +4 -4
  15. package/docs/operator-manual-vietnamese.md +21 -41
  16. package/docs/publishing-for-teams.md +5 -0
  17. package/docs/quality-benchmark.md +179 -23
  18. package/docs/release-install-policy.md +6 -6
  19. package/docs/runtime-harness-standard.md +3 -3
  20. package/docs/security-threat-model.md +1 -1
  21. package/docs/subagent-orchestration-capabilities.md +19 -51
  22. package/docs/subagents-and-multiagent.md +21 -47
  23. package/docs/task-implementation-contract.md +3 -3
  24. package/docs/task-lifecycle-vietnamese.md +1 -1
  25. package/docs/team-onboarding.md +2 -2
  26. package/docs/vi/architecture.md +26 -13
  27. package/docs-site/architecture.html +1 -1
  28. package/docs-site/benchmark.html +89 -95
  29. package/docs-site/commands.html +1 -1
  30. package/docs-site/content/benchmark.html +88 -94
  31. package/docs-site/content/en/benchmark.html +134 -90
  32. package/docs-site/content/en/index.html +19 -19
  33. package/docs-site/content/en/runtime.html +4 -4
  34. package/docs-site/content/en/whats-new.html +12 -1
  35. package/docs-site/content/en/workflows.html +2 -2
  36. package/docs-site/content/index.html +19 -19
  37. package/docs-site/content/runtime.html +4 -4
  38. package/docs-site/content/whats-new.html +12 -1
  39. package/docs-site/content/workflows.html +2 -2
  40. package/docs-site/context-engine.html +1 -1
  41. package/docs-site/en/architecture.html +1 -1
  42. package/docs-site/en/benchmark.html +135 -91
  43. package/docs-site/en/commands.html +1 -1
  44. package/docs-site/en/context-engine.html +1 -1
  45. package/docs-site/en/index.html +20 -20
  46. package/docs-site/en/mcp-approval.html +1 -1
  47. package/docs-site/en/mcp-auth.html +1 -1
  48. package/docs-site/en/mcp-servers.html +1 -1
  49. package/docs-site/en/mcp.html +1 -1
  50. package/docs-site/en/permissions.html +1 -1
  51. package/docs-site/en/profiles.html +1 -1
  52. package/docs-site/en/quickstart.html +1 -1
  53. package/docs-site/en/runtime.html +5 -5
  54. package/docs-site/en/scope.html +1 -1
  55. package/docs-site/en/security.html +1 -1
  56. package/docs-site/en/team.html +1 -1
  57. package/docs-site/en/whats-new.html +13 -2
  58. package/docs-site/en/workflows.html +3 -3
  59. package/docs-site/index.html +20 -20
  60. package/docs-site/mcp-approval.html +1 -1
  61. package/docs-site/mcp-auth.html +1 -1
  62. package/docs-site/mcp-servers.html +1 -1
  63. package/docs-site/mcp.html +1 -1
  64. package/docs-site/permissions.html +1 -1
  65. package/docs-site/profiles.html +1 -1
  66. package/docs-site/quickstart.html +1 -1
  67. package/docs-site/runtime.html +5 -5
  68. package/docs-site/scope.html +1 -1
  69. package/docs-site/security.html +1 -1
  70. package/docs-site/team.html +1 -1
  71. package/docs-site/whats-new.html +13 -2
  72. package/docs-site/workflows.html +3 -3
  73. package/evals/fixtures/benchmark-assurance-evidence.invalid.json +1 -1
  74. package/evals/fixtures/benchmark-assurance-evidence.valid.json +1 -1
  75. package/evals/fixtures/codex-relative-efficiency-v1.invalid.json +4 -0
  76. package/evals/fixtures/codex-relative-efficiency-v1.valid.json +94 -0
  77. package/evals/fixtures/helper-request.invalid.json +1 -1
  78. package/evals/fixtures/helper-request.valid.json +1 -1
  79. package/evals/fixtures/piagent-webui/subagent-tree-v1.valid.json +7 -0
  80. package/evals/fixtures/task-authority-snapshot.valid.json +4 -4
  81. package/evals/fs4-readiness-matrix.v1.json +2 -2
  82. package/evals/fs5-causal-arm.v1.json +2 -2
  83. package/evals/fs5-pilot-protocol.v1.json +3 -3
  84. package/evals/fs5-pilot-protocol.v2.json +8 -8
  85. package/evals/fs5-pilot-protocol.v3.json +8 -8
  86. package/evals/fs5-pilot-protocol.v4.json +6 -6
  87. package/evals/fs5-pilot-protocol.v5.json +7 -6
  88. package/evals/private-holdout-v1/public-exposure.v1.json +1 -1
  89. package/evals/runtime-conformance-v1/README.md +43 -0
  90. package/evals/runtime-conformance-v1/evaluate-paid-canary.mjs +198 -0
  91. package/evals/runtime-conformance-v1/lane.json +57 -0
  92. package/evals/runtime-conformance-v1/paid-canary.json +31 -0
  93. package/evals/runtime-conformance-v1/runner.mjs +456 -0
  94. package/package.json +1 -1
  95. package/packages/piagent-core/README.md +13 -4
  96. package/packages/piagent-core/benchmark/benchmark-codex-relative-efficiency.js +698 -0
  97. package/packages/piagent-core/benchmark/benchmark-core.js +4 -3
  98. package/packages/piagent-core/benchmark/benchmark-forensics.js +4 -2
  99. package/packages/piagent-core/benchmark/benchmark-preflight.js +3 -1
  100. package/packages/piagent-core/benchmark/benchmark-provider-free-evidence.js +345 -0
  101. package/packages/piagent-core/benchmark/benchmark-record-validation.js +74 -5
  102. package/packages/piagent-core/benchmark/benchmark-report.js +15 -1
  103. package/packages/piagent-core/benchmark/benchmark-runtime-causal.js +244 -0
  104. package/packages/piagent-core/benchmark/benchmark-stage-diagnostic.js +201 -4
  105. package/packages/piagent-core/benchmark/benchmark-suite.js +22 -1
  106. package/packages/piagent-core/benchmark/benchmark-summary-support.js +4 -0
  107. package/packages/piagent-core/benchmark/benchmark-usage.js +15 -4
  108. package/packages/piagent-core/capabilities/runtime-integrity.js +3 -0
  109. package/packages/piagent-core/extensions/acceptance-behavior-proof.js +806 -0
  110. package/packages/piagent-core/extensions/acceptance-receipt.js +89 -48
  111. package/packages/piagent-core/extensions/context-engine.js +1 -11
  112. package/packages/piagent-core/extensions/context-index-policy.js +4 -0
  113. package/packages/piagent-core/extensions/context-query-paths.js +57 -0
  114. package/packages/piagent-core/extensions/criterion-context-pack.js +16 -3
  115. package/packages/piagent-core/extensions/criterion-graph.js +4 -1
  116. package/packages/piagent-core/extensions/fallback-base-policy.ts +101 -0
  117. package/packages/piagent-core/extensions/guard-types.ts +4 -0
  118. package/packages/piagent-core/extensions/piagent-guard.ts +194 -226
  119. package/packages/piagent-core/extensions/task-lifecycle.js +32 -0
  120. package/packages/piagent-core/extensions/task-state.js +69 -8
  121. package/packages/piagent-core/extensions/verification-intelligence.js +49 -12
  122. package/packages/piagent-core/extensions/workspace-evidence-roots.js +52 -16
  123. package/packages/piagent-core/package.json +1 -1
  124. package/packages/piagent-core/policies/base-policy.json +42 -2
  125. package/packages/piagent-core/policy/authority-manifest.v1.json +3 -3
  126. package/packages/piagent-core/prompts/be-to-fe.md +2 -8
  127. package/packages/piagent-core/prompts/plan.md +1 -1
  128. package/packages/piagent-core/prompts/platform-improve.md +2 -6
  129. package/packages/piagent-core/prompts/review.md +2 -7
  130. package/packages/piagent-core/prompts/scout.md +1 -1
  131. package/packages/piagent-core/prompts/task.md +4 -3
  132. package/packages/piagent-core/runtime/hooks/completion-hook.ts +7 -6
  133. package/packages/piagent-core/runtime/hooks/input-hook.ts +6 -0
  134. package/packages/piagent-core/runtime/hooks/session-start-hook.ts +100 -1
  135. package/packages/piagent-core/runtime/hooks/tool-result-hook.ts +58 -72
  136. package/packages/piagent-core/runtime/hooks/tool-result-value-helpers.ts +66 -0
  137. package/packages/piagent-core/runtime/inspection/bounded-emission-guard.ts +286 -0
  138. package/packages/piagent-core/runtime/inspection/runtime-event-store.ts +37 -7
  139. package/packages/piagent-core/runtime/inspection/task-subagent-tree.ts +27 -4
  140. package/packages/piagent-core/runtime/inspection/tool-failure-classification.ts +20 -3
  141. package/packages/piagent-core/runtime/inspection/webui-snapshot.ts +47 -11
  142. package/packages/piagent-core/runtime/orchestration/helper-lifecycle.ts +16 -12
  143. package/packages/piagent-core/runtime/orchestration/minimal-delegation-policy.ts +71 -0
  144. package/packages/piagent-core/runtime/orchestration/owned-work-budget.ts +28 -20
  145. package/packages/piagent-core/runtime/orchestration/role-policy.ts +27 -9
  146. package/packages/piagent-core/runtime/orchestration/subagent-tool-policy.ts +329 -0
  147. package/packages/piagent-core/runtime/policy/block-remedy.ts +0 -4
  148. package/packages/piagent-core/runtime/product/efficiency-metrics.ts +3 -0
  149. package/packages/piagent-core/runtime/quality/edit-freshness-guard.ts +102 -0
  150. package/packages/piagent-core/runtime/quality/exact-output-contract.ts +20 -10
  151. package/packages/piagent-core/runtime/quality/performance-assurance.ts +3 -3
  152. package/packages/piagent-core/runtime/recovery/continuation-budget.ts +1 -1
  153. package/packages/piagent-core/runtime/recovery/handoff-projection.ts +4 -1
  154. package/packages/piagent-core/runtime/recovery/pristine-task-policy-refresh.ts +91 -0
  155. package/packages/piagent-core/runtime/recovery/recovery-policy.ts +7 -4
  156. package/packages/piagent-core/runtime/recovery/resume-state.ts +112 -14
  157. package/packages/piagent-core/runtime/recovery/semantic-repair-handshake.ts +7 -13
  158. package/packages/piagent-core/runtime/recovery/task-evidence-coverage-refresh.ts +217 -0
  159. package/packages/piagent-core/runtime/registration/context-commands.ts +16 -2
  160. package/packages/piagent-core/runtime/registration/knowledge-tools.ts +9 -5
  161. package/packages/piagent-core/runtime/registration/operator-request-intake.ts +49 -0
  162. package/packages/piagent-core/runtime/registration/policy-tools.ts +4 -4
  163. package/packages/piagent-core/runtime/registration/session-commands.ts +1 -1
  164. package/packages/piagent-core/runtime/registration/task-completion-tools.ts +12 -9
  165. package/packages/piagent-core/runtime/registration/task-evidence-tools.ts +21 -12
  166. package/packages/piagent-core/runtime/registration/task-start-guidance.ts +19 -13
  167. package/packages/piagent-core/runtime/registration/task-start-manifest.ts +54 -0
  168. package/packages/piagent-core/runtime/registration/task-start-tool.ts +57 -57
  169. package/packages/piagent-core/runtime/runtime-limits.ts +16 -2
  170. package/packages/piagent-core/runtime/session/adaptive-context-analysis.ts +461 -0
  171. package/packages/piagent-core/runtime/session/adaptive-context-governor.ts +464 -0
  172. package/packages/piagent-core/runtime/session/adaptive-context-ledger.ts +265 -0
  173. package/packages/piagent-core/runtime/session/message-signals.ts +7 -3
  174. package/packages/piagent-core/runtime/session/operator-request-carry.ts +44 -0
  175. package/packages/piagent-core/runtime/session/runtime-state.ts +11 -7
  176. package/packages/piagent-core/runtime/session/source-checkout-read-grants.ts +35 -0
  177. package/packages/piagent-core/runtime/session/system-prompt.ts +76 -14
  178. package/packages/piagent-core/runtime/session/task-contract-presentation.ts +39 -0
  179. package/packages/piagent-core/runtime/tools/phase-tools.ts +8 -8
  180. package/packages/piagent-core/runtime/tools/tool-groups.ts +12 -2
  181. package/packages/piagent-core/runtime/verification/exact-verifier-reuse.ts +101 -0
  182. package/packages/piagent-core/runtime/workflows/runtime-intake-compaction.ts +164 -0
  183. package/packages/piagent-core/runtime/workflows/task-intake.ts +167 -39
  184. package/packages/piagent-core/subagents/piagent-oracle.md +5 -3
  185. package/packages/piagent-core/subagents/piagent-planner.md +6 -4
  186. package/packages/piagent-core/subagents/piagent-reviewer.md +5 -3
  187. package/packages/piagent-core/subagents/piagent-scout.md +4 -2
  188. package/packages/piagent-core/subagents/piagent-worker.md +6 -4
  189. package/packages/piagent-webui/dist/client/assets/{MarkdownMessage-CxeZqTte.js → MarkdownMessage-B4XehTCU.js} +1 -1
  190. package/packages/piagent-webui/dist/client/assets/{index-DynIMe8I.js → index-CgOEyRPG.js} +4 -4
  191. package/packages/piagent-webui/dist/client/index.html +1 -1
  192. package/packages/piagent-webui/extension/piagent-webui.ts +3 -1
  193. package/packages/piagent-webui/extension/same-session-bridge.ts +1 -1
  194. package/packages/piagent-webui/extension/session-stream-adapter.ts +33 -10
  195. package/packages/piagent-webui/gateway/gateway-session-stream.ts +37 -21
  196. package/packages/piagent-webui/gateway/session-operation-lifecycle.ts +237 -0
  197. package/packages/piagent-webui/gateway/session-operation-retry-control.ts +38 -0
  198. package/packages/piagent-webui/gateway/session-operation-watchdog.ts +9 -2
  199. package/packages/piagent-webui/gateway/session-runtime-supervisor.ts +8 -7
  200. package/packages/piagent-webui/ownership/session-lease-store.ts +5 -2
  201. package/packages/piagent-webui/server/core-inspection-provider.ts +2 -1
  202. package/packages/piagent-webui/server/transcript-projection.ts +91 -5
  203. package/schemas/codex-relative-efficiency-v1.schema.json +155 -0
  204. package/schemas/piagent-webui/runtime-event-v2.schema.json +15 -5
  205. package/schemas/piagent-webui/subagent-tree-v1.schema.json +11 -1
  206. package/schemas/project-profile.schema.json +1 -1
  207. package/schemas/task-contract.schema.json +15 -0
  208. package/scripts/benchmark-runner-core.mjs +14 -6
  209. package/scripts/benchmark-runner-finalization.mjs +4 -0
  210. package/scripts/configure-subagents.sh +42 -34
  211. package/scripts/helper-evaluation.mjs +1 -1
  212. package/scripts/team-doctor.sh +1 -1
  213. package/scripts/verify-local.sh +5 -0
@@ -11,11 +11,11 @@ export function registerKnowledgeTools(pi: ExtensionAPI, deps: Record<string, an
11
11
  registerPiagentTool(pi, {
12
12
  name: "piagent_orchestration_policy",
13
13
  label: "Piagent Orchestration Policy",
14
- description: "Return solo-first subagent, review lens, model-role, and Field Guide policy for the current project.",
15
- promptSnippet: "Use this before planning medium/large tasks so orchestration stays single-agent-first and token-aware.",
14
+ description: "Return parent-direct helper limits, review lenses, model-role, and Field Guide policy for the current project.",
15
+ promptSnippet: "Use this before planning medium/large tasks so the parent stays direct and helper use remains a measured token optimization.",
16
16
  promptGuidelines: [
17
- "Default to the parent agent plus bounded subagents only when they reduce context risk or improve review quality.",
18
- "Use review lenses instead of spawning a broad swarm.",
17
+ "Default to the parent model working directly; review quality alone does not justify another model run.",
18
+ "Use at most one fresh read-only helper only with two independent lanes and at least 30% projected total token savings; never delegate implementation or retry a deterministic failure.",
19
19
  "Treat Field Guide memory as advisory and verify it against current repository files."
20
20
  ],
21
21
  parameters: Type.Object({
@@ -45,7 +45,9 @@ export function registerKnowledgeTools(pi: ExtensionAPI, deps: Record<string, an
45
45
  path: fieldGuidePath,
46
46
  exists: fieldGuideExists
47
47
  },
48
- stance: "single-agent-first; subagents are opt-in tools for bounded scout, planning, and review"
48
+ helperBudget: "1-readonly-total/0-retries/0-writers",
49
+ minimumProjectedNetSaving: 0.3,
50
+ stance: "parent-direct; one fresh read-only helper only when runtime-measured net token savings justify it"
49
51
  };
50
52
  const text = params.detail === "full"
51
53
  ? JSON.stringify(payload, null, 2)
@@ -53,6 +55,8 @@ export function registerKnowledgeTools(pi: ExtensionAPI, deps: Record<string, an
53
55
  `mode: ${payload.defaultMode}`,
54
56
  `maxConcurrentSubagents: ${payload.maxConcurrentSubagents}`,
55
57
  `helpersMode: ${payload.helpersMode}`,
58
+ `helperBudget: ${payload.helperBudget}`,
59
+ `minimumProjectedNetSaving: ${Math.round(payload.minimumProjectedNetSaving * 100)}%`,
56
60
  `reviewLenses: ${payload.defaultReviewLenses.join(", ")}`,
57
61
  `fieldGuide: ${payload.fieldGuide.enabled ? `${payload.fieldGuide.path} (${payload.fieldGuide.exists ? "exists" : "missing"})` : "off"}`,
58
62
  `fieldGuidePolicy: ${payload.fieldGuide.writePolicy}, maxLines=${payload.fieldGuide.maxLines}`,
@@ -0,0 +1,49 @@
1
+ import { OPERATOR_REQUEST_MAX_CHARS } from "../../extensions/task-state.js";
2
+
3
+ type SessionContext = {
4
+ sessionManager: {
5
+ getBranch?: () => unknown;
6
+ getEntries?: () => unknown;
7
+ };
8
+ };
9
+
10
+ function userMessageText(message: unknown): string {
11
+ if (!message || typeof message !== "object" || (message as Record<string, unknown>).role !== "user") return "";
12
+ const content = (message as Record<string, unknown>).content;
13
+ if (typeof content === "string") return content;
14
+ if (!Array.isArray(content)) return "";
15
+ // The host stores the exact operator prose in the first text block. Later
16
+ // text blocks may contain extracted attachment bodies and must not be copied
17
+ // into private task truth as if the operator typed them.
18
+ const first = content.find((block) => block && typeof block === "object"
19
+ && (block as Record<string, unknown>).type === "text"
20
+ && typeof (block as Record<string, unknown>).text === "string") as Record<string, unknown> | undefined;
21
+ return typeof first?.text === "string" ? first.text : "";
22
+ }
23
+
24
+ export function latestOperatorRequest(ctx: SessionContext): string | undefined {
25
+ let entries: unknown[] = [];
26
+ try {
27
+ const branch = ctx.sessionManager.getBranch?.();
28
+ entries = Array.isArray(branch) ? branch : [];
29
+ if (entries.length === 0) {
30
+ const all = ctx.sessionManager.getEntries?.();
31
+ entries = Array.isArray(all) ? all : [];
32
+ }
33
+ } catch {
34
+ return undefined;
35
+ }
36
+ for (let index = entries.length - 1; index >= 0; index -= 1) {
37
+ const entry = entries[index] as Record<string, any> | undefined;
38
+ if (entry?.type !== "message") continue;
39
+ const text = userMessageText(entry.message);
40
+ if (text.trim()) return text;
41
+ }
42
+ return undefined;
43
+ }
44
+
45
+ export function boundedOperatorRequest(value: unknown, redactText: (value: string) => string): string | undefined {
46
+ if (typeof value !== "string" || !value.trim()) return undefined;
47
+ const redacted = redactText(value);
48
+ return Array.from(redacted).length <= OPERATOR_REQUEST_MAX_CHARS ? redacted : undefined;
49
+ }
@@ -30,7 +30,7 @@ export function registerPolicyTools(pi: ExtensionAPI, deps: Record<string, any>)
30
30
  "When recovery is necessary, load only the smallest group that resolves the reported missing evidence."
31
31
  ],
32
32
  parameters: Type.Object({
33
- groups: Type.Array(StringEnum(["intake", "governance", "task", "recovery", "policy", "retrieval", "knowledge", "onboarding", "usage"] as const), { minItems: 1 })
33
+ groups: Type.Array(StringEnum(["intake", "governance", "task", "recovery", "policy", "retrieval", "source", "knowledge", "onboarding", "usage"] as const), { minItems: 1 })
34
34
  }),
35
35
  async execute(_toolCallId, params, _signal, _onUpdate, ctx) {
36
36
  const groups = [...new Set(params.groups)] as PiagentToolGroup[];
@@ -429,8 +429,8 @@ export function registerPolicyTools(pi: ExtensionAPI, deps: Record<string, any>)
429
429
  registerPiagentTool(pi, {
430
430
  name: "piagent_task_gate_check",
431
431
  label: "Piagent Task Gate Check",
432
- description: "Check whether a governed task has enough context, verify evidence, and trace before claiming done.",
433
- promptSnippet: "Use this before final on source-changing tasks.",
432
+ description: "Check whether a governed task has enough context, current-tree exact verifier evidence, and trace before claiming done or rerunning verification.",
433
+ promptSnippet: "Use before final and before repeating an exact verifier; run only commands listed as missing for the current tree.",
434
434
  parameters: Type.Object({
435
435
  taskId: Type.String({ minLength: 1 }),
436
436
  changedFiles: Type.Optional(Type.Array(Type.String()))
@@ -451,7 +451,7 @@ export function registerPolicyTools(pi: ExtensionAPI, deps: Record<string, any>)
451
451
  ...verifierCommandInstructions(result.missingVerifyCommands),
452
452
  `warnings: ${result.warnings.join("; ") || "none"}`
453
453
  ].join("\n");
454
- return { content: [{ type: "text", text }], details: { ...result, task: projected } };
454
+ return { content: [{ type: "text", text }], details: { ...result, task: projected ? { taskId: projected.taskId, taskRunId: projected.taskRunId } : undefined } };
455
455
  }
456
456
  });
457
457
 
@@ -334,7 +334,7 @@ export function registerSessionCommands(pi: ExtensionAPI, deps: Record<string, a
334
334
  const approvals = feature ? [feature.externalAction ? "external-action" : "", feature.destructiveAction ? "destructive-action" : "", feature.permissionExpansion ? "permission-expansion" : ""].filter(Boolean) : [];
335
335
  const modelRoute = feature ? await evaluateModelRoute?.(ctx, feature, snapshot) : undefined;
336
336
  const retrievalRoute = feature ? await evaluateRetrievalRoute?.(ctx, feature) : undefined;
337
- return { runtime: snapshot, modelRoute: modelRoute?.status === "ok" ? modelRoute.decision : null, retrievalRoute: retrievalRoute ?? null, scope: task?.scope ?? [], protectedPaths: effectiveProtectedPaths(policy, profile).readProtectedPaths, activeToolGroups: pi.getActiveTools(), helperMode: helpersMode(), helperBudget: "2-concurrent/3-total", executionBackend: "host", executionBoundary: "host execution is not a sandbox", approvals, blockers: feature?.protectedTarget ? ["protected-target"] : [], controlMode: evaluation.status === "ok" && evaluation.decision.mode === "shadow" ? "shadow" : resolveRuntimePolicy(profile).finalGate === "enforce" ? "enforce" : "assist" };
337
+ return { runtime: snapshot, modelRoute: modelRoute?.status === "ok" ? modelRoute.decision : null, retrievalRoute: retrievalRoute ?? null, scope: task?.scope ?? [], protectedPaths: effectiveProtectedPaths(policy, profile).readProtectedPaths, activeToolGroups: pi.getActiveTools(), helperMode: helpersMode(), helperBudget: "1-readonly-total/0-retries/0-writers", executionBackend: "host", executionBoundary: "host execution is not a sandbox", approvals, blockers: feature?.protectedTarget ? ["protected-target"] : [], controlMode: evaluation.status === "ok" && evaluation.decision.mode === "shadow" ? "shadow" : resolveRuntimePolicy(profile).finalGate === "enforce" ? "enforce" : "assist" };
338
338
  }
339
339
  });
340
340
 
@@ -9,6 +9,9 @@ import { readTrajectoryStore } from "../trajectory/trajectory-store.ts";
9
9
 
10
10
  type TaskContract = any;
11
11
 
12
+ function taskErrorDetails(task: TaskContract, reasonCode: string): Record<string, unknown> {
13
+ return { reasonCode, taskId: task.taskId, taskRunId: task.taskRunId, outcome: task.trace?.outcome, attempt: task.attempt };
14
+ }
12
15
 
13
16
  export function registerTaskCompletionTools(pi: ExtensionAPI, deps: Record<string, any>): void {
14
17
  const {
@@ -41,7 +44,7 @@ export function registerTaskCompletionTools(pi: ExtensionAPI, deps: Record<strin
41
44
  return { content: [{ type: "text", text: `Task not found: ${params.taskId}` }], isError: true };
42
45
  }
43
46
  if (task.trace.outcome !== "pending") {
44
- return { content: [{ type: "text", text: `Task ${task.taskId} is immutable after ${task.trace.outcome}; context evidence was not changed.` }], details: task, isError: true };
47
+ return { content: [{ type: "text", text: `Task ${task.taskId} is immutable after ${task.trace.outcome}; context evidence was not changed.` }], details: taskErrorDetails(task, "task-immutable"), isError: true };
45
48
  }
46
49
  const activeIdentity = runtimeState.taskIdentity(ctx);
47
50
  if (!activeIdentity || activeIdentity.taskId !== task.taskId || activeIdentity.taskRunId !== task.taskRunId) {
@@ -127,7 +130,7 @@ export function registerTaskCompletionTools(pi: ExtensionAPI, deps: Record<strin
127
130
  return { content: [{ type: "text", text: `Task not found: ${params.taskId}` }], isError: true };
128
131
  }
129
132
  if (task.trace.outcome !== "pending") {
130
- return { content: [{ type: "text", text: `Task ${task.taskId} is immutable after ${task.trace.outcome}; verify evidence was not changed.` }], details: task, isError: true };
133
+ return { content: [{ type: "text", text: `Task ${task.taskId} is immutable after ${task.trace.outcome}; verify evidence was not changed.` }], details: taskErrorDetails(task, "task-immutable"), isError: true };
131
134
  }
132
135
 
133
136
  const observedEntries = [
@@ -239,7 +242,7 @@ export function registerTaskCompletionTools(pi: ExtensionAPI, deps: Record<strin
239
242
  return { content: [{ type: "text", text: `Task not found: ${params.taskId}` }], isError: true };
240
243
  }
241
244
  if (task.trace.outcome !== "pending") {
242
- return { content: [{ type: "text", text: `Task ${task.taskId} is immutable after ${task.trace.outcome}; memory evidence was not changed.` }], details: task, isError: true };
245
+ return { content: [{ type: "text", text: `Task ${task.taskId} is immutable after ${task.trace.outcome}; memory evidence was not changed.` }], details: taskErrorDetails(task, "task-immutable"), isError: true };
243
246
  }
244
247
 
245
248
  const safeFiles = params.files.map((file) => ({
@@ -257,7 +260,7 @@ export function registerTaskCompletionTools(pi: ExtensionAPI, deps: Record<strin
257
260
 
258
261
  return {
259
262
  content: [{ type: "text", text: `Memory citations recorded for ${task.taskId}: ${params.files.length} file(s)` }],
260
- details: task
263
+ details: compactTaskDetails(task)
261
264
  };
262
265
  }
263
266
  });
@@ -284,13 +287,13 @@ export function registerTaskCompletionTools(pi: ExtensionAPI, deps: Record<strin
284
287
  return { content: [{ type: "text", text: `Task not found: ${params.taskId}` }], isError: true };
285
288
  }
286
289
  if (task.trace.outcome !== "pending") {
287
- return { content: [{ type: "text", text: `Task ${task.taskId} is immutable after ${task.trace.outcome}; its final trace was not replaced.` }], details: task, isError: true };
290
+ return { content: [{ type: "text", text: `Task ${task.taskId} is immutable after ${task.trace.outcome}; its final trace was not replaced.` }], details: taskErrorDetails(task, "task-immutable"), isError: true };
288
291
  }
289
292
  if (params.outcome !== "completed" && !params.friction?.trim() && !task.failureReason?.trim()) {
290
- return { content: [{ type: "text", text: `Trace ${params.outcome} requires a concrete friction/reason so the next attempt does not repeat the same work.` }], details: task, isError: true };
293
+ return { content: [{ type: "text", text: `Trace ${params.outcome} requires a concrete friction/reason so the next attempt does not repeat the same work.` }], details: taskErrorDetails(task, "trace-friction-required"), isError: true };
291
294
  }
292
295
  if (params.outcome === "failed" && !params.failedAt && !task.failedAt) {
293
- return { content: [{ type: "text", text: "A failed trace requires failedAt to identify the lifecycle phase." }], details: task, isError: true };
296
+ return { content: [{ type: "text", text: "A failed trace requires failedAt to identify the lifecycle phase." }], details: taskErrorDetails(task, "trace-failure-phase-required"), isError: true };
294
297
  }
295
298
 
296
299
  const rawChangedFiles = params.changedFiles ?? task.changedFiles;
@@ -363,7 +366,7 @@ export function registerTaskCompletionTools(pi: ExtensionAPI, deps: Record<strin
363
366
  ...verifierCommandInstructions(gate.missingVerifyCommands)
364
367
  ].join("\n")
365
368
  }],
366
- details: { gate, task: nextTask },
369
+ details: { gate, task: taskErrorDetails(nextTask, "completion-gate-blocked") },
367
370
  isError: true
368
371
  };
369
372
  }
@@ -426,7 +429,7 @@ export function registerTaskCompletionTools(pi: ExtensionAPI, deps: Record<strin
426
429
 
427
430
  return {
428
431
  content: [{ type: "text", text: `Trace recorded for ${nextTask.taskId}: ${params.outcome}${gate.decision === "fail" ? ` (gate warning: missing ${gate.missing.join(", ")})` : ""}` }],
429
- details: { task: written, gate, completionReceipt: buildCompletionReceiptView(written, { cwd: ctx.cwd, gate }) }
432
+ details: { task: compactTaskDetails(written), gate, completionReceipt: buildCompletionReceiptView(written, { cwd: ctx.cwd, gate }) }
430
433
  };
431
434
  }
432
435
  });
@@ -2,6 +2,9 @@ import type { ExtensionAPI } from "@earendil-works/pi-coding-agent";
2
2
 
3
3
  type WorkPlanStep = any;
4
4
 
5
+ function taskErrorDetails(task: any, reasonCode: string): Record<string, unknown> {
6
+ return { reasonCode, taskId: task.taskId, taskRunId: task.taskRunId, outcome: task.trace?.outcome, attempt: task.attempt };
7
+ }
5
8
 
6
9
  export function registerTaskEvidenceTools(pi: ExtensionAPI, deps: Record<string, any>): void {
7
10
  const {
@@ -9,7 +12,7 @@ export function registerTaskEvidenceTools(pi: ExtensionAPI, deps: Record<string,
9
12
  compactTaskDetails, effectiveProtectedPaths, extensionDir, extractDocument, loadProfileFromContext,
10
13
  matchesAnyPath, matchesProtectedPath, nowIso, path, permissionOverrideFromContext,
11
14
  policy, readTask, recordTaskProgressCheckpoints, redactText, registerPiagentTool,
12
- resolveDocumentPath, resolveDocumentRoots, resolvePermissionProfile, safeTaskId, verifyProjectCapabilityState,
15
+ resolveDocumentPath, resolveDocumentRoots, resolvePermissionProfile, runtimeState, safeTaskId, verifyProjectCapabilityState,
13
16
  writeTask
14
17
  } = deps;
15
18
  registerPiagentTool(pi, {
@@ -37,24 +40,26 @@ export function registerTaskEvidenceTools(pi: ExtensionAPI, deps: Record<string,
37
40
  return { content: [{ type: "text", text: `Task not found in this session: ${params.taskId}` }], isError: true };
38
41
  }
39
42
  if (task.trace.outcome !== "pending") {
40
- return { content: [{ type: "text", text: `Task ${task.taskId} is immutable after ${task.trace.outcome}; start a fresh session for another attempt.` }], details: task, isError: true };
43
+ return { content: [{ type: "text", text: `Task ${task.taskId} is immutable after ${task.trace.outcome}; start a fresh session for another attempt.` }], details: taskErrorDetails(task, "task-immutable"), isError: true };
41
44
  }
42
45
  const stepId = safeTaskId(params.stepId).slice(0, 40);
43
46
  const step = task.workPlan.find((item) => item.id === stepId);
44
47
  if (!step) {
45
- const valid = task.workPlan.map((item) => `${item.id} (${item.status})`).join(", ");
48
+ const validSteps = task.workPlan.slice(0, 20).map((item) => ({ id: item.id, status: item.status }));
49
+ const valid = validSteps.map((item) => `${item.id} (${item.status})`).join(", ");
46
50
  const actionable = task.workPlan
47
51
  .filter((item) => item.status === "in-progress" || (
48
52
  item.status === "pending"
49
53
  && (item.dependsOn ?? []).every((dependency) => task.workPlan.some((candidate) => candidate.id === dependency && ["done", "skipped"].includes(candidate.status)))
50
54
  ))
51
- .map((item) => item.id);
55
+ .map((item) => item.id)
56
+ .slice(0, 20);
52
57
  return {
53
58
  content: [{
54
59
  type: "text",
55
60
  text: `Work-plan step not found: ${stepId}. Valid step IDs: ${valid || "none"}. Currently actionable: ${actionable.join(", ") || "none"}.`
56
61
  }],
57
- details: task.workPlan,
62
+ details: { ...taskErrorDetails(task, "work-plan-step-not-found"), stepId, validSteps, actionable },
58
63
  isError: true
59
64
  };
60
65
  }
@@ -66,7 +71,7 @@ export function registerTaskEvidenceTools(pi: ExtensionAPI, deps: Record<string,
66
71
  if ((params.status === "in-progress" || params.status === "done") && unresolved.length > 0) {
67
72
  return {
68
73
  content: [{ type: "text", text: `Step ${stepId} is blocked by unfinished dependencies: ${unresolved.join(", ")}` }],
69
- details: { step, unresolved },
74
+ details: { ...taskErrorDetails(task, "work-plan-step-blocked"), stepId, unresolved: unresolved.slice(0, 20) },
70
75
  isError: true
71
76
  };
72
77
  }
@@ -75,13 +80,13 @@ export function registerTaskEvidenceTools(pi: ExtensionAPI, deps: Record<string,
75
80
  return { content: [{ type: "text", text: `A concrete note is required when step ${stepId} fails.` }], isError: true };
76
81
  }
77
82
  if ((step.status === "done" || step.status === "skipped") && params.status !== step.status) {
78
- return { content: [{ type: "text", text: `Work-plan step ${stepId} is already ${step.status} and cannot be reopened in the same attempt.` }], details: step, isError: true };
83
+ return { content: [{ type: "text", text: `Work-plan step ${stepId} is already ${step.status} and cannot be reopened in the same attempt.` }], details: { ...taskErrorDetails(task, "work-plan-step-immutable"), stepId, status: step.status }, isError: true };
79
84
  }
80
85
  if (step.status === params.status) {
81
- return { content: [{ type: "text", text: `Work-plan step ${stepId} is already ${params.status}; no state change was recorded.` }], details: step, isError: true };
86
+ return { content: [{ type: "text", text: `Work-plan step ${stepId} is already ${params.status}; no state change was recorded.` }], details: { ...taskErrorDetails(task, "work-plan-step-unchanged"), stepId, status: step.status }, isError: true };
82
87
  }
83
88
  if (step.status === "failed" && params.status === "in-progress" && !note) {
84
- return { content: [{ type: "text", text: `A concrete note is required to reopen failed step ${stepId} within this attempt.` }], details: step, isError: true };
89
+ return { content: [{ type: "text", text: `A concrete note is required to reopen failed step ${stepId} within this attempt.` }], details: { ...taskErrorDetails(task, "work-plan-step-reopen-note-required"), stepId, status: step.status }, isError: true };
85
90
  }
86
91
 
87
92
  const recordedAt = nowIso();
@@ -254,7 +259,7 @@ export function registerTaskEvidenceTools(pi: ExtensionAPI, deps: Record<string,
254
259
  registerPiagentTool(pi, {
255
260
  name: "piagent_source_checkout",
256
261
  label: "Piagent Source Checkout",
257
- description: "Cache and refresh an external Git repository for targeted local inspection.",
262
+ description: "Prepare an external Git repository for bounded, read-only local inspection.",
258
263
  promptSnippet: "Use this before reading a user-provided external source repository.",
259
264
  promptGuidelines: [
260
265
  "Use for GitHub/GitLab/Bitbucket source repositories supplied by the user.",
@@ -264,15 +269,19 @@ export function registerTaskEvidenceTools(pi: ExtensionAPI, deps: Record<string,
264
269
  repoRef: Type.String({ minLength: 3, description: "owner/repo, host/owner/repo, https URL, or git@host:owner/repo.git" }),
265
270
  forceUpdate: Type.Optional(Type.Boolean({ description: "Fetch immediately even if the cache was refreshed recently." }))
266
271
  }),
267
- async execute(_toolCallId, params) {
272
+ async execute(_toolCallId, params, _signal, _onUpdate, ctx) {
268
273
  try {
269
274
  const repo = checkoutReferenceRepo(params.repoRef, params.forceUpdate === true);
275
+ const grantedPath = runtimeState.grantSourceCheckoutReadRoot(ctx, repo.checkoutPath);
276
+ repo.checkoutPath = grantedPath;
270
277
  const text = [
271
278
  "Source cache ready:",
272
279
  `path: ${repo.checkoutPath}`,
273
280
  `url: ${repo.cloneUrl}`,
274
281
  `commit: ${repo.commit ?? "unknown"}`,
275
- `fetched: ${repo.fetched ? "yes" : "no"}`
282
+ `fetched: ${repo.fetched ? "yes" : "no"}`,
283
+ "access: read-only for this session via read, grep, find, or ls",
284
+ "Do not run a checkout script or use shell mutation against this shared cache."
276
285
  ].join("\n");
277
286
  return {
278
287
  content: [{ type: "text", text }],
@@ -102,12 +102,18 @@ export function criticalProofSection(
102
102
  acceptanceTests: FocusedTestCandidate[],
103
103
  criterionProofs: CriterionProofGuidance[]
104
104
  ): string[] {
105
- if (proofGuidance.length === 0) return [];
105
+ if (proofGuidance.length === 0 && criterionProofs.length === 0) return [];
106
106
  const candidateMap = new Map<string, string[]>();
107
107
  for (const candidate of acceptanceTests) {
108
108
  candidateMap.set(candidate.criterionId, [...new Set([...(candidateMap.get(candidate.criterionId) ?? []), candidate.path])]);
109
109
  }
110
110
  let mappedCandidate = false;
111
+ const criterionTag = (criterionId: string): string => {
112
+ const candidates = candidateMap.get(criterionId) ?? [];
113
+ if (candidates.length === 0) return `[${criterionId}:fallback]`;
114
+ mappedCandidate = true;
115
+ return `[${criterionId}:candidate=${candidates.join("|")}]`;
116
+ };
111
117
  const proofLines = proofGuidance.flatMap((proof) => {
112
118
  const criterionIds = [...new Set(criterionProofs
113
119
  .filter((entry) => entry.guidance.includes(proof))
@@ -115,21 +121,19 @@ export function criticalProofSection(
115
121
  if (criterionIds.length === 0) {
116
122
  return [`- [fallback] ${proof}`];
117
123
  }
118
- const tags = criterionIds.map((criterionId) => {
119
- const candidates = candidateMap.get(criterionId) ?? [];
120
- if (candidates.length === 0) {
121
- return `[${criterionId}:fallback]`;
122
- }
123
- mappedCandidate = true;
124
- return `[${criterionId}:candidate=${candidates.join("|")}]`;
125
- });
124
+ const tags = criterionIds.map(criterionTag);
126
125
  return [`- ${tags.join("")} ${proof}`];
127
126
  });
127
+ const genericTags = criterionProofs.filter((entry) => entry.guidance.length === 0).map((entry) => criterionTag(entry.criterionId));
128
+ const genericLines = genericTags.length > 0
129
+ ? [`- ${genericTags.join("")} Prove every tagged observable clause with linked live assertions; the generic verifier alone is insufficient.`]
130
+ : [];
128
131
  return [[
129
132
  "Critical behavioral proof:",
130
- ...(!mappedCandidate ? ["No concrete criterion-linked focused test was selected."] : []),
133
+ ...(!mappedCandidate && genericTags.length === 0 ? ["No concrete criterion-linked focused test was selected."] : []),
131
134
  ...proofLines,
132
- "Candidate tags are locations, not proof; add/update durable scoped tests with live assertions and changed-test evidence. Each fallback tag requires adding/updating a durable scoped focused test with live assertions. If test scope is unavailable, report missing durable proof and do not claim completion. Run proof after the criterion's final intended mutation and before exact verifiers; rerun after later target mutation or a runtime-authorized same-tree infrastructure retry. Transient/print-only probes and prose are insufficient."
135
+ ...genericLines,
136
+ "Candidate tags are locations, not proof; add/update durable scoped tests with live assertions and changed-test evidence. Each fallback tag requires adding/updating a durable scoped focused test with live assertions. If test scope is unavailable, report missing durable proof and do not claim completion. Run proof after the criterion's final intended mutation and before exact verifiers; rerun after later target mutation or a runtime-authorized same-tree infrastructure retry. Transient/print-only probes and prose are insufficient. Claims must match diff/tests; never claim unproved behavior."
133
137
  ].join("\n")];
134
138
  }
135
139
 
@@ -150,7 +154,8 @@ export function taskCriticalProofSection(
150
154
  repositoryFiles: string[],
151
155
  acceptanceProofGuidance: (taskOrText: unknown) => string[],
152
156
  isAcceptanceTestPath: (file: string) => boolean,
153
- acceptanceLanguageAdapterForPath: (file: string) => { disposition?: string }
157
+ acceptanceLanguageAdapterForPath: (file: string) => { disposition?: string },
158
+ requireGenericFallback = true
154
159
  ): string[] {
155
160
  const proofGuidance = task.changeMode === "source-change" && task.mutationPolicy !== "forbidden"
156
161
  ? acceptanceProofGuidance(task)
@@ -158,6 +163,7 @@ export function taskCriticalProofSection(
158
163
  const acceptanceTests = concretePlannedAcceptanceTests(
159
164
  plannedContext, repositoryFiles, task.criterionGraph, isAcceptanceTestPath, acceptanceLanguageAdapterForPath
160
165
  );
161
- const criterionProofs = behavioralCriterionProofGuidance(task.criterionGraph, acceptanceProofGuidance);
166
+ const criterionProofs = behavioralCriterionProofGuidance(task.criterionGraph, acceptanceProofGuidance)
167
+ .filter((entry) => requireGenericFallback || entry.guidance.length > 0);
162
168
  return criticalProofSection(proofGuidance, acceptanceTests, criterionProofs);
163
169
  }
@@ -0,0 +1,54 @@
1
+ type RepositoryManifestCompatibilityReason = "missing-details-provider" | "nonfunction-details-provider" | "invalid-details-result";
2
+
3
+ type TaskStartRepositoryManifest = {
4
+ files: string[];
5
+ complete: boolean;
6
+ candidateCount: number;
7
+ provider: "details" | "legacy-array-fallback";
8
+ compatibilityReason?: RepositoryManifestCompatibilityReason;
9
+ };
10
+
11
+ function normalizedManifestFiles(value: unknown): string[] {
12
+ if (!Array.isArray(value)) return [];
13
+ return [...new Set(value.filter((item): item is string => typeof item === "string" && item.trim().length > 0))];
14
+ }
15
+
16
+ /**
17
+ * Keep rolling upgrades safe when a long-lived host has the legacy extension
18
+ * entrypoint cached but imports a newer task-start registration module. The
19
+ * legacy array provider cannot prove whether its result was truncated, so the
20
+ * compatibility path deliberately reports `complete: false` and records the
21
+ * provider/reason in the durable task trace.
22
+ */
23
+ export function resolveTaskStartRepositoryManifestProvider(deps: Record<string, any>): {
24
+ read(cwd: string, maxFiles?: number): TaskStartRepositoryManifest;
25
+ } {
26
+ const legacyProvider = deps.repositoryFileManifest;
27
+ if (typeof legacyProvider !== "function") {
28
+ throw new TypeError("Piagent task-start registration requires repositoryFileManifest to be a function.");
29
+ }
30
+ const detailsProvider = deps.repositoryFileManifestDetails;
31
+ const compatibilityReason: RepositoryManifestCompatibilityReason | undefined = detailsProvider === undefined
32
+ ? "missing-details-provider"
33
+ : typeof detailsProvider !== "function"
34
+ ? "nonfunction-details-provider"
35
+ : undefined;
36
+ const legacyManifest = (cwd: string, maxFiles?: number,
37
+ reason: RepositoryManifestCompatibilityReason = compatibilityReason ?? "invalid-details-result"): TaskStartRepositoryManifest => {
38
+ const files = normalizedManifestFiles(legacyProvider(cwd, maxFiles));
39
+ return { files, complete: false, candidateCount: files.length, provider: "legacy-array-fallback", compatibilityReason: reason };
40
+ };
41
+ if (compatibilityReason) return { read: legacyManifest };
42
+ return {
43
+ read(cwd: string, maxFiles?: number): TaskStartRepositoryManifest {
44
+ const details = detailsProvider(cwd, maxFiles);
45
+ if (!details || typeof details !== "object" || !Array.isArray(details.files)) {
46
+ return legacyManifest(cwd, maxFiles, "invalid-details-result");
47
+ }
48
+ const files = normalizedManifestFiles(details.files);
49
+ return { files, complete: details.complete === true,
50
+ candidateCount: Number.isInteger(details.candidateCount) ? Math.max(files.length, details.candidateCount) : files.length,
51
+ provider: "details" };
52
+ }
53
+ };
54
+ }