mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-11-07 07:55:47 +01:00
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Harness selection maps supported agents to Paperclip Runner. > - The native Runner already supports Grok Build, but its migration mapping excludes Grok. > - Grok agents also have a thinking-effort setting that the native execution contract does not carry. > - This pull request adds the migration path and preserves explicit Grok effort through startup and resume. > - Operators can upgrade an existing Grok agent without silently losing its model, account, or effort. ## Linked Issues or Issue Description **What happened?** Existing Grok Build agents have no Runner selector or upgrade action. Adding only the mapping would route conversion through the Codex fallback and lose Grok execution semantics. Native execution also ignores the legacy reasoning-effort field. **Expected behavior** A qualified Grok agent can explicitly upgrade. The saved model, account, instructions, history, and supported effort remain intact. Unsupported legacy settings fail before saving. The provider must acknowledge the selected effort before inference. **Steps to reproduce** 1. Open the runtime page of an existing Grok Build agent on a qualified Linux x64 or macOS ARM64 computer. 2. Observe that the Runner selector and upgrade action are absent. 3. Compare the native execution input with the saved legacy reasoning-effort setting. Related work: #13882 introduced native Grok Build. #15422 covers the earlier default-runner work. This change extends the current shared selection contract. ## What Changed - Add the Grok harness mapping, qualified platform limits, upgrade action, and company-scoped computer qualification. - Translate Grok settings to ACPX without changing the selected account or explicit model. Resolve the legacy default model to the native default. Preserve the effective legacy search policy. Reject incompatible command, argument, permission, and turn-limit settings. - Add native execution input v10 for explicit Grok effort and search policy. Preserve historical schemas. Enforce disabled search through Grok requirements and bind it across startup and recovery. - Apply and verify ACP `reasoning_effort` before prompting. Bind the value to provider descriptors, checkpoints, runtime identity, and resume checks. - Use the Grok catalog and effort choices in the task composer. Apply task model and effort overrides while retaining the agent account and search policy. Update runtime controls, portability/default handling, documentation, and capability maps. - Advance Pi and Copilot source attestations for changed shared transport and recovery code. Keep historical identities intact and update all Rust integration fixtures. ## Verification - Migration and review regressions were reproduced before the fixes. Coverage includes the Grok mapping, default model, explicit effort, implicit search policy, task overrides, and incompatible legacy settings. The broad migration, route, immutable-input, and composer check passed 478 tests. The final catalog and task-input check passed 135 tests. - Runner startup, recovery, search policy, and profile checks passed 361 tests, with one platform skip. The final source-identity, search-admission, and transport event-pump check passed 89 tests. - All 757 Rust workspace tests passed for the broad implementation. After the final source attestation update, all 55 ACPX profile/session tests passed. - `pnpm -r typecheck` and `pnpm build` passed for the implementation. Server and Runner typecheck passed after the final production change. Server `tsc --noEmit` passed after the final test-fixture change. Docker context integrity, token gates, maps, generated protocol data, and replay goldens passed. - The v10 startup fixture initially used an ambient local Grok login and failed on CI. An explicitly empty credential home reproduced all four failures. The fixture now supplies a synthetic key and a private empty home. A separate case checks that missing credentials fail closed. All seven selected Grok v10 and Claude v9 startup cases pass without host credentials. - The full local suite could not complete because embedded PostgreSQL failed before database-backed fixture setup. The broad executor-suite attempt hit the same limit and was stopped. Neither attempt is counted as passing. Current-head CI must verify those paths before merge. - Earlier CI found missing Rust input fields, stale source attestations and exact fixture pins, a missing Docker context entry, and an outdated import expectation. Those failures were fixed and checked locally. The CI logs retain the original failures. - Different earlier CI runs also failed unchanged cleanup pagination, device-login timeout, and chat lost-response recovery cases. Later attempts passed the cleanup and login cases without changes to those implementations. Their causes remain unproven. The original evidence remains in [the first run](https://github.com/paperclipai/paperclip/actions/runs/40563402840), [the login run](https://github.com/paperclipai/paperclip/actions/runs/40571128328), and [the chat run](https://github.com/paperclipai/paperclip/actions/runs/40573072163). A passing retry does not establish a root-cause fix. - Both reviews rate current head `16fba9b4cdf12bc7945e6628912e79da42959f35` 5/5. All current-head CI gates pass in [run 40576580699](https://github.com/paperclipai/paperclip/actions/runs/40576580699). All 60 check runs completed cleanly, including the four intentional skips and the neutral 5/5 Paperclip review. The legacy security status is successful. No review threads remain unresolved. - Staging migration and a real shell task are pending deployment. This PR does not claim full Grok provider qualification. Reviewers can verify the operator path by upgrading a qualified Grok agent, checking its saved model and effort, and running a shell task through native execution. ## Risks - Explicit effort and search policy require the matching Runner and sidecar build. Older immutable execution inputs keep their existing schemas. - Providers that omit or reject the selected effort fail before a prompt. Unsupported effort does not fall back silently. - Existing agents retain their saved runner until an explicit change. New Grok agents use native execution only on qualified platforms. - Grok token and cost reporting remain unverified, as before. ## Model Used OpenAI Codex, GPT-6 family. The exact serving model ID and context-window capacity are not exposed in this session. Used reasoning, source editing, shell execution, and browser inspection. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes: #` / `Refs: #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and the focused suites pass; the full local suite has the environment limitation recorded above - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing>
575 lines
18 KiB
TypeScript
575 lines
18 KiB
TypeScript
import type { NativeRunRuntimeGrant } from "./runtime-context.js";
|
|
import type {
|
|
PrpEvent,
|
|
PrpStructuredRunResult,
|
|
} from "../protocol/replay-contract.js";
|
|
import type { NativeSessionCapabilities, NativeTurnControlCapabilities, NativeUserMessage } from "./types.js";
|
|
import {
|
|
PAPERCLIP_RUNTIME_REQUEST_SCHEMA_V2,
|
|
parsePaperclipQuestionResponse,
|
|
type PaperclipQuestionResponse,
|
|
type PaperclipQuestionSet,
|
|
type PaperclipRuntimeRequestOrigin,
|
|
} from "./question-set.js";
|
|
|
|
export * from "./question-set.js";
|
|
|
|
export const HARNESS_DRIVER_CONTRACT_VERSION = 1 as const;
|
|
|
|
export interface HarnessDriverConfigIssue {
|
|
path: string;
|
|
code: string;
|
|
message: string;
|
|
}
|
|
|
|
export type HarnessDriverConfigValidation =
|
|
| { ok: true; config: Record<string, unknown>; issues: [] }
|
|
| { ok: false; config: null; issues: HarnessDriverConfigIssue[] };
|
|
|
|
export interface HarnessTranscriptSnapshot {
|
|
schema: "paperclip-runner/harness-transcript/v1";
|
|
complete: boolean;
|
|
eventCount: number;
|
|
events: PrpEvent[];
|
|
omissionReason: string | null;
|
|
}
|
|
|
|
export interface HarnessDriverDescriptor {
|
|
kind: string;
|
|
displayName: string;
|
|
version: string;
|
|
protocolVersion?: string;
|
|
capabilities: NativeSessionCapabilities;
|
|
runtimeContextCapabilities?: NativeRuntimeContextCapabilities;
|
|
}
|
|
|
|
export interface NativeRuntimeContextCapabilities {
|
|
instructions: "native" | "unsupported";
|
|
skills: "native" | "unsupported";
|
|
mcp: "native" | "unsupported";
|
|
}
|
|
|
|
export interface OpenHarnessSessionInput {
|
|
runId: string;
|
|
normalizedSessionId: string;
|
|
workingDirectory: string;
|
|
/** Abort provider bootstrap and release any not-yet-returned provider state. */
|
|
signal?: AbortSignal;
|
|
}
|
|
|
|
export interface HarnessSessionRecoveryOptions {
|
|
normalizeReplayPayload?(payload: PrpEvent["payload"]): PrpEvent["payload"];
|
|
/** Abort provider recovery and release any not-yet-returned provider state. */
|
|
signal: AbortSignal;
|
|
}
|
|
|
|
export class HarnessCapabilityUnavailableError extends Error {
|
|
readonly operation: string;
|
|
|
|
constructor(operation: string, detail: string) {
|
|
super(`${operation} is unavailable: ${detail}`);
|
|
this.name = "HarnessCapabilityUnavailableError";
|
|
this.operation = operation;
|
|
}
|
|
}
|
|
|
|
export class HarnessReconciliationError extends Error {
|
|
readonly recoverable = true;
|
|
|
|
constructor(message: string) {
|
|
super(message);
|
|
this.name = "HarnessReconciliationError";
|
|
}
|
|
}
|
|
|
|
export class HarnessOperationAlreadyTerminalError extends Error {
|
|
readonly code = "already_terminal" as const;
|
|
|
|
constructor(operation: string) {
|
|
super(`${operation} lost a race with the committed turn terminal`);
|
|
this.name = "HarnessOperationAlreadyTerminalError";
|
|
}
|
|
}
|
|
|
|
export class HarnessStaleTurnError extends Error {
|
|
readonly code = "stale_turn" as const;
|
|
|
|
constructor(turnId: string) {
|
|
super(`turn ${turnId} is not the active turn`);
|
|
this.name = "HarnessStaleTurnError";
|
|
}
|
|
}
|
|
|
|
export type HarnessRuntimeRequestKind =
|
|
| "command_approval"
|
|
| "file_approval"
|
|
| "permission_approval"
|
|
| "user_input"
|
|
| "elicitation";
|
|
|
|
export interface HarnessRuntimeRequest {
|
|
requestId: string;
|
|
requestKind: HarnessRuntimeRequestKind;
|
|
method: string;
|
|
turnId: string;
|
|
itemId: string;
|
|
status: "pending";
|
|
prompt: string;
|
|
details: Record<string, unknown>;
|
|
/** Provider-neutral presentation model for structured input requests. */
|
|
input?: PaperclipQuestionSet;
|
|
/** Diagnostic origin only; provider response shapes never enter PRP. */
|
|
origin?: PaperclipRuntimeRequestOrigin;
|
|
}
|
|
|
|
export type HarnessRuntimeRequestResolution =
|
|
| { action: "accept" | "accept_for_session" | "decline" | "cancel" }
|
|
| {
|
|
action: "submit";
|
|
answers: Record<string, { answers: string[] }>;
|
|
}
|
|
| {
|
|
action: "submit";
|
|
content: Record<string, unknown>;
|
|
}
|
|
| {
|
|
action: "submit";
|
|
response: PaperclipQuestionResponse;
|
|
};
|
|
|
|
export type HarnessRuntimeRequestAction =
|
|
HarnessRuntimeRequestResolution["action"];
|
|
|
|
export type HarnessRuntimeRequestHandoffResult =
|
|
"handed_off" | "already_settled";
|
|
|
|
/**
|
|
* A runtime-input handoff commits its durable state transition before the
|
|
* method returns. Provider interruption is cleanup only: it remains observed,
|
|
* but cannot acquire mutation authority or reverse an already committed turn.
|
|
*/
|
|
export interface HarnessRuntimeRequestHandoff {
|
|
result: HarnessRuntimeRequestHandoffResult;
|
|
cleanup: Promise<void>;
|
|
}
|
|
|
|
const RUNTIME_REQUEST_ACTIONS: readonly HarnessRuntimeRequestAction[] = [
|
|
"accept",
|
|
"accept_for_session",
|
|
"decline",
|
|
"cancel",
|
|
"submit",
|
|
];
|
|
|
|
export class HarnessRuntimeRequestResolutionError extends Error {
|
|
readonly code = "invalid_resolution" as const;
|
|
readonly requestKind: HarnessRuntimeRequestKind;
|
|
|
|
constructor(requestKind: HarnessRuntimeRequestKind, detail: string) {
|
|
super(`${requestKind} rejected its resolution: ${detail}`);
|
|
this.name = "HarnessRuntimeRequestResolutionError";
|
|
this.requestKind = requestKind;
|
|
}
|
|
}
|
|
|
|
function plainRecord(value: unknown): Record<string, unknown> | null {
|
|
return typeof value === "object" && value !== null && !Array.isArray(value)
|
|
? (value as Record<string, unknown>)
|
|
: null;
|
|
}
|
|
|
|
function parseAnswers(
|
|
value: unknown,
|
|
): Record<string, { answers: string[] }> | null {
|
|
const fields = plainRecord(value);
|
|
if (fields === null || Object.keys(fields).length === 0) return null;
|
|
const parsed: Record<string, { answers: string[] }> = {};
|
|
for (const [field, entry] of Object.entries(fields)) {
|
|
const answer = plainRecord(entry);
|
|
if (answer === null || !Array.isArray(answer.answers)) return null;
|
|
if (answer.answers.some((value) => typeof value !== "string")) return null;
|
|
parsed[field] = { answers: [...(answer.answers as string[])] };
|
|
}
|
|
return parsed;
|
|
}
|
|
|
|
/**
|
|
* Validates a resolution against the kind of request it answers, so a
|
|
* mismatched submit fails closed instead of degrading into an accepted empty
|
|
* response. Every driver runs this before it touches its provider, and the
|
|
* browser transport runs it again at its own untrusted edge.
|
|
*/
|
|
export function parseHarnessRuntimeRequestResolution(
|
|
requestKind: HarnessRuntimeRequestKind,
|
|
value: unknown,
|
|
questionSet?: PaperclipQuestionSet,
|
|
): HarnessRuntimeRequestResolution {
|
|
const candidate = plainRecord(value) ?? {};
|
|
const rawAction = candidate.action;
|
|
if (
|
|
typeof rawAction !== "string" ||
|
|
!RUNTIME_REQUEST_ACTIONS.includes(rawAction as HarnessRuntimeRequestAction)
|
|
) {
|
|
throw new HarnessRuntimeRequestResolutionError(
|
|
requestKind,
|
|
`unsupported action ${JSON.stringify(rawAction) ?? "undefined"}`,
|
|
);
|
|
}
|
|
const action = rawAction as HarnessRuntimeRequestAction;
|
|
if (
|
|
action !== "submit" &&
|
|
("answers" in candidate ||
|
|
"content" in candidate ||
|
|
"response" in candidate)
|
|
) {
|
|
throw new HarnessRuntimeRequestResolutionError(
|
|
requestKind,
|
|
`${action} does not carry submitted form data`,
|
|
);
|
|
}
|
|
|
|
if (action === "submit" && "response" in candidate) {
|
|
if (requestKind !== "user_input" && requestKind !== "elicitation") {
|
|
throw new HarnessRuntimeRequestResolutionError(
|
|
requestKind,
|
|
"approval requests do not accept submitted question responses",
|
|
);
|
|
}
|
|
if ("answers" in candidate || "content" in candidate) {
|
|
throw new HarnessRuntimeRequestResolutionError(
|
|
requestKind,
|
|
"canonical submissions cannot also carry provider-specific answers or content",
|
|
);
|
|
}
|
|
if (questionSet === undefined) {
|
|
throw new HarnessRuntimeRequestResolutionError(
|
|
requestKind,
|
|
"canonical submission requires the persisted question set",
|
|
);
|
|
}
|
|
try {
|
|
return {
|
|
action,
|
|
response: parsePaperclipQuestionResponse(
|
|
questionSet,
|
|
candidate.response,
|
|
),
|
|
};
|
|
} catch (error) {
|
|
throw new HarnessRuntimeRequestResolutionError(
|
|
requestKind,
|
|
error instanceof Error ? error.message : "invalid question response",
|
|
);
|
|
}
|
|
}
|
|
|
|
if (requestKind === "user_input") {
|
|
if (action === "accept" || action === "accept_for_session") {
|
|
throw new HarnessRuntimeRequestResolutionError(
|
|
requestKind,
|
|
"user input requires submit, decline, or cancel",
|
|
);
|
|
}
|
|
if (action !== "submit") return { action };
|
|
if ("content" in candidate) {
|
|
throw new HarnessRuntimeRequestResolutionError(
|
|
requestKind,
|
|
"user input submissions carry answers, not content",
|
|
);
|
|
}
|
|
const answers = parseAnswers(candidate.answers);
|
|
if (answers === null) {
|
|
throw new HarnessRuntimeRequestResolutionError(
|
|
requestKind,
|
|
"submit requires answers shaped as { field: { answers: [string] } }",
|
|
);
|
|
}
|
|
return { action, answers };
|
|
}
|
|
|
|
if (requestKind === "elicitation") {
|
|
if (action === "accept_for_session") {
|
|
throw new HarnessRuntimeRequestResolutionError(
|
|
requestKind,
|
|
"elicitation does not support session acceptance",
|
|
);
|
|
}
|
|
if (action !== "submit") return { action };
|
|
if ("answers" in candidate) {
|
|
throw new HarnessRuntimeRequestResolutionError(
|
|
requestKind,
|
|
"elicitation submissions carry content, not answers",
|
|
);
|
|
}
|
|
const content = plainRecord(candidate.content);
|
|
if (content === null || Object.keys(content).length === 0) {
|
|
throw new HarnessRuntimeRequestResolutionError(
|
|
requestKind,
|
|
"submit requires a non-empty content object",
|
|
);
|
|
}
|
|
return { action, content: structuredClone(content) };
|
|
}
|
|
|
|
if (action === "submit") {
|
|
throw new HarnessRuntimeRequestResolutionError(
|
|
requestKind,
|
|
"approval requests do not accept submitted form data",
|
|
);
|
|
}
|
|
return { action };
|
|
}
|
|
|
|
/**
|
|
* The single payload shape every driver emits for a terminal
|
|
* `runtime_request.*` fact. Request identity travels in the payload as well as
|
|
* the event binding, so a consumer that only reads payloads still settles the
|
|
* request it belongs to.
|
|
*/
|
|
export type HarnessRuntimeRequestOutcome = {
|
|
requestId: string;
|
|
requestKind: HarnessRuntimeRequestKind;
|
|
turnId: string;
|
|
itemId: string;
|
|
action?: HarnessRuntimeRequestAction;
|
|
reason?: string;
|
|
/** Enables content-free lifecycle counters partitioned by adapter. */
|
|
adapter?: string;
|
|
requestType?: "input" | "permission";
|
|
/** Canonical submitted answers retained for durable replay and audit UI. */
|
|
response?: PaperclipQuestionResponse;
|
|
};
|
|
|
|
export function harnessRuntimeRequestOutcome(
|
|
request: Pick<
|
|
HarnessRuntimeRequest,
|
|
"requestId" | "requestKind" | "turnId" | "itemId" | "origin" | "input"
|
|
>,
|
|
outcome: {
|
|
action?: HarnessRuntimeRequestAction | null;
|
|
reason?: string | null;
|
|
response?: PaperclipQuestionResponse | null;
|
|
} = {},
|
|
): HarnessRuntimeRequestOutcome {
|
|
return {
|
|
requestId: request.requestId,
|
|
requestKind: request.requestKind,
|
|
turnId: request.turnId,
|
|
itemId: request.itemId,
|
|
...(outcome.action ? { action: outcome.action } : {}),
|
|
...(outcome.reason ? { reason: outcome.reason } : {}),
|
|
...(outcome.response
|
|
? { response: structuredClone(outcome.response) }
|
|
: {}),
|
|
...(request.input
|
|
? {
|
|
...(request.origin?.adapter
|
|
? { adapter: request.origin.adapter }
|
|
: {}),
|
|
requestType: "input" as const,
|
|
}
|
|
: {}),
|
|
};
|
|
}
|
|
|
|
/** Canonical non-replayable expiration used when a live input moves to a durable wait. */
|
|
export function harnessRuntimeInputExpiredOutcome(
|
|
request: HarnessRuntimeRequest,
|
|
reason: "durable_handoff" | "provider_process_lost",
|
|
): Omit<HarnessRuntimeRequestOutcome, "requestKind"> & {
|
|
requestKind: "runtime";
|
|
replayAllowed: false;
|
|
request: Record<string, unknown>;
|
|
} {
|
|
return {
|
|
...harnessRuntimeRequestOutcome(request, { reason }),
|
|
requestKind: "runtime",
|
|
replayAllowed: false,
|
|
request: {
|
|
schema: PAPERCLIP_RUNTIME_REQUEST_SCHEMA_V2,
|
|
requestKind: "runtime",
|
|
requestId: request.requestId,
|
|
type: "input",
|
|
status: request.status,
|
|
prompt: request.prompt,
|
|
input: structuredClone(request.input),
|
|
origin: structuredClone(request.origin),
|
|
turnId: request.turnId,
|
|
itemId: request.itemId,
|
|
},
|
|
};
|
|
}
|
|
|
|
export interface HarnessThreadGoal {
|
|
threadId: string;
|
|
objective: string;
|
|
status: "active" | "paused" | "blocked" | "limited" | "usageLimited" | "budgetLimited" | "complete";
|
|
tokenBudget: number | null;
|
|
tokensUsed: number;
|
|
timeUsedSeconds: number;
|
|
createdAt: number;
|
|
updatedAt: number;
|
|
}
|
|
|
|
export type HarnessGoalOperation =
|
|
| { action: "get"; requestId?: string }
|
|
| {
|
|
action: "set";
|
|
objective: string;
|
|
tokenBudget?: number | null;
|
|
status?: HarnessThreadGoal["status"];
|
|
requestId?: string;
|
|
}
|
|
| { action: "pause" | "resume" | "clear"; requestId?: string };
|
|
|
|
export interface HarnessThreadLineageEntry {
|
|
threadId: string;
|
|
providerSessionId: string | null;
|
|
parentThreadId: string | null;
|
|
depth: number;
|
|
nickname: string | null;
|
|
role: string | null;
|
|
status: string;
|
|
}
|
|
|
|
export interface PersistedHarnessSemanticResult {
|
|
result: PrpStructuredRunResult;
|
|
fingerprint: string;
|
|
callId?: string | null;
|
|
turnId: string;
|
|
}
|
|
|
|
export interface PersistedHarnessTurnTerminal {
|
|
turnId: string;
|
|
fingerprint: string;
|
|
}
|
|
|
|
export interface AcpxSessionIdentity {
|
|
kind: "acpx";
|
|
normalizedSessionId: string;
|
|
acpxRecordId: string;
|
|
backendSessionId: string;
|
|
agentSessionId: string;
|
|
profileDigest: string;
|
|
workspaceDigest: string;
|
|
requestedModel: string;
|
|
effectiveModel: string;
|
|
/** Missing on legacy snapshots; those used the historical approve-reads behavior. */
|
|
permissionMode?: "approve-all" | "approve-paperclip" | "approve-reads" | "deny-all";
|
|
/** Effective provider mode identifier, bound to the session identity. */
|
|
mode?: string;
|
|
piThinkingLevel?: "off" | "low" | "high" | "max";
|
|
claudeEffort?: "low" | "medium" | "high" | "xhigh" | "max";
|
|
grokEffort?: "minimal" | "low" | "medium" | "high" | "xhigh";
|
|
grokDisableWebSearch?: boolean;
|
|
providerLifetimeFenceCandidates: readonly [number, number, number];
|
|
}
|
|
|
|
export type PersistedHarnessProviderIdentity = AcpxSessionIdentity;
|
|
|
|
export interface PersistedHarnessSession {
|
|
driverKind: string;
|
|
/** Execution-host startup root, revalidated before a cold provider launch. */
|
|
workingDirectory?: string;
|
|
driverSessionId: string;
|
|
providerSessionId?: string | null;
|
|
runId?: string;
|
|
normalizedSessionId?: string;
|
|
activeTurnId?: string | null;
|
|
semanticResult?: PersistedHarnessSemanticResult | null;
|
|
terminalTurns?: PersistedHarnessTurnTerminal[];
|
|
codexUsageBaseline?: { baseline: Record<string, number>; latest: Record<string, number> };
|
|
/** A result-less terminal task may spend this fail-closed one-shot recovery allowance. */
|
|
dispositionOnlyRecoveryConsumed?: boolean;
|
|
/** Exact accepted provider turn that spent the disposition-only allowance. */
|
|
dispositionOnlyRecoveryTurnId?: string | null;
|
|
pendingRuntimeRequests?: HarnessRuntimeRequest[];
|
|
goal?: HarnessThreadGoal | null;
|
|
lineage?: HarnessThreadLineageEntry[];
|
|
lastSourceSequence?: number;
|
|
/** Tagged provider identity used to reject cross-profile recovery. */
|
|
providerIdentity?: PersistedHarnessProviderIdentity;
|
|
/** Narrow escape hatch for a durable response-wake when a provider cannot reload its prior native session. */
|
|
providerRecoveryPolicy?:
|
|
| "same_session_only"
|
|
| "allow_replacement_after_governed_wait"
|
|
| "allow_replacement_after_resume_failure";
|
|
}
|
|
|
|
export interface HarnessSessionRecoveryResult {
|
|
recovered: boolean;
|
|
session?: HarnessSession;
|
|
reason?: string;
|
|
}
|
|
|
|
export interface HarnessSession {
|
|
supportsTurnReasoning?(): boolean;
|
|
turnControlCapabilities?(): NativeTurnControlCapabilities | null;
|
|
ids(): {
|
|
driverSessionId: string;
|
|
providerSessionId?: string | null;
|
|
displayId?: string | null;
|
|
};
|
|
events(): AsyncIterable<PrpEvent>;
|
|
attachRun?(input: { runId: string; currentRunGrant?: NativeRunRuntimeGrant }): Promise<void> | void;
|
|
startTurn(input: {
|
|
message: NativeUserMessage;
|
|
/** Set by orchestration only after successful provider-session recovery. */
|
|
continuation?: true;
|
|
requestedCollaborationMode?: "default" | "plan";
|
|
/** OpenCode/OpenRouter only. Applies to this turn, never subsequent turns. */
|
|
reasoningMode?: "default" | "disabled";
|
|
}): Promise<{
|
|
turnId: string;
|
|
effectiveCollaborationMode?: "default" | "plan";
|
|
}>;
|
|
steer?(input: {
|
|
/** Queued follow-ups remain distinct from active-turn steering. */
|
|
mode?: "steer" | "follow_up";
|
|
turnId: string;
|
|
message: NativeUserMessage;
|
|
correlationId?: string;
|
|
}): Promise<void>;
|
|
interrupt?(input: {
|
|
turnId?: string;
|
|
reason?: string;
|
|
signal?: AbortSignal;
|
|
}): Promise<void>;
|
|
pendingRuntimeRequests?(): HarnessRuntimeRequest[];
|
|
resolveRuntimeRequest?(input: {
|
|
/** Durable controller command identity, distinct from the native request identity. */
|
|
commandId?: string;
|
|
requestId: string;
|
|
turnId: string;
|
|
resolution: HarnessRuntimeRequestResolution;
|
|
}): Promise<void>;
|
|
handoffRuntimeRequest?(input: {
|
|
requestId: string;
|
|
turnId: string;
|
|
reason: "durable_handoff";
|
|
/** Do not synchronously commit if runtime ownership is already revoked. */
|
|
signal: AbortSignal;
|
|
}): HarnessRuntimeRequestHandoff;
|
|
goal?(input: HarnessGoalOperation): Promise<HarnessThreadGoal | null>;
|
|
lineage?(): HarnessThreadLineageEntry[];
|
|
read?(): Promise<Record<string, unknown>>;
|
|
reconcile?(): Promise<Record<string, unknown>>;
|
|
usage?(): Promise<Record<string, unknown> | null>;
|
|
transcript?(): Promise<HarnessTranscriptSnapshot>;
|
|
snapshot(): Promise<PersistedHarnessSession>;
|
|
/** Relinquish controller authority without semantically closing the session. */
|
|
detachControllerForRestart?(): Promise<void>;
|
|
close(input: { reason: string; force?: boolean }): Promise<void>;
|
|
}
|
|
|
|
/** A local harness implementation hidden behind the runner daemon boundary. */
|
|
export interface HarnessDriver {
|
|
descriptor(): Promise<HarnessDriverDescriptor>;
|
|
validateConfig?(config: unknown): Promise<HarnessDriverConfigValidation>;
|
|
openSession(input: OpenHarnessSessionInput): Promise<HarnessSession>;
|
|
recoverSession?(
|
|
snapshot: PersistedHarnessSession,
|
|
options: HarnessSessionRecoveryOptions,
|
|
): Promise<HarnessSessionRecoveryResult>;
|
|
}
|