Files
PaperClipAI/patches/acpx@0.12.0.patch
T
DottaandPaperclip 3b4b270650 fix(adapters): preserve ACP terminal failure diagnostics (#14573)
## Thinking Path

> - Paperclip is the open source app people use to manage AI agents for
work.
> - The shared ACP adapter engine records agent failures for operators.
> - ACP providers can report a failure category, title, and detailed
cause.
> - Our patch kept only the category in the saved error, so an operator
could not diagnose a failure when tracing was off.
> - This pull request preserves redacted provider diagnostics in the run
error, transcript, and structured run result.
> - Operators can now inspect the provider message and any supplied
request ID or stack trace after the run ends.

## Linked Issues or Issue Description

Refs #13889 (the diagnostic gap; this PR does not update the bundled
Claude version).
Refs #14484 (related model-refusal classification; this PR retains
diagnostics for all terminal failure categories).

**What happened?**
An ACP turn failed with only `ACP agent reported a terminal service
failure.` The provider's title and details were available in memory but
absent from the saved error and transcript.

**Expected behavior**
The run retains useful provider diagnostics even when raw tracing is
disabled. Credentials remain redacted. A size limit must report
truncation instead of silently removing the cause.

**Steps to reproduce**
1. Run an ACP agent that returns an error-severity typed session
failure.
2. Include an HTTP error, request ID, and stack text in its title and
details.
3. Inspect the failed run with tracing disabled. Before this change,
only the category survives.

## What Changed

- Both pinned ACPX patches pass complete error text to the in-memory
callback, so redaction happens before truncation.
- The shared engine retains the sanitized category, title, and details
in `resultJson.terminalSessionFailure` and includes the text in the run
error and error transcript.
- Diagnostics redact configured environment values even under arbitrary
names, unknown launch-environment values, connection URL passwords, run
credentials, and common credential syntax. Known boolean settings remain
readable, while credential values are redacted even when embedded in
other text. Diagnostics remove control characters and invalid Unicode.
- Title and detail limits keep escaped transcript JSON below the
server's chunk limit. Truncated fields include an omission count. The
safe run-result projection preserves a byte-bounded diagnostic preview
when the result exceeds its byte budget, with an explicit pointer to the
full adapter-bounded run error and transcript.
- The existing UI and CLI display the error. Diagnostics do not become
assistant output. Issue continuation summaries and session-compaction
prompts receive only the generic category, preventing provider text from
becoming handoff instructions. Existing quota classification, warnings,
timeout precedence, and control-channel failure precedence remain in
place.
- Regression tests cover real ACP child processes with both pinned
versions in one-shot and persistent modes, credential redaction, request
IDs after the old 4 KiB cutoff, transcript parsing, storage bounds, and
database retrieval of oversized multibyte diagnostics.

## Verification

- Full CI on `20ad4f5f1f66c46d2c260e6ad0339cbea607b4cf`: **54 passed, 2
intentionally skipped, no pending or failing checks**. Includes
typechecking, build, all Vitest shards, Runner checks, browser E2E, and
the canary packaging/public-install dry run.
- Greptile: **5/5** on this commit. Superagent security scan passes. All
review threads are resolved.
- Local verification passed: shared ACP engine suite (395 tests); real
Claude ACP child-process and diagnostic regressions across both pinned
runtimes and both execution modes; run retrieval and model-handoff
regressions (59 tests); ACPX patch packaging (16 tests); full typecheck
and build. Affected package typechecks and focused tests were rerun
after review fixes.
- The broad local `pnpm test:run` was stopped after review edits made
its cached imports stale. Fresh targeted runs pass, including both
affected server suites. Cold-build import failures were also rerun after
dependency builds: chat integration (1,063 tests) and tool access (351
tests) pass. The final commit's complete CI matrix is green.

## Risks

- Provider diagnostic text is untrusted. This change retains more of it
in company-scoped run records. Redaction and size bounds apply before
persistence.
- Diagnostics are limited to fields the provider supplies. Old runs
cannot recover discarded error text.
- No schema migration, recovery-policy change, or new Telemetry or
OpenTelemetry export.

## Model Used

- OpenAI GPT-6 through Codex, with reasoning, repository inspection,
code editing, and test execution. The exact serving model ID and
context-window size are not exposed in this session.

## 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 (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [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>
2026-09-29 10:18:30 -05:00

240 lines
10 KiB
Diff

diff --git a/dist/live-checkpoint-ClPCSdrW.js b/dist/live-checkpoint-ClPCSdrW.js
index 243c9d13bcba520923b63adfddad75cf2d94362d..e2a254d4e57fb2e246ccb742f2e9de5a8a461d20 100644
--- a/dist/live-checkpoint-ClPCSdrW.js
+++ b/dist/live-checkpoint-ClPCSdrW.js
@@ -1532,7 +1532,7 @@ const ZED_TAG_KEYS = /* @__PURE__ */ new Set([
"RedactedThinking",
"ToolUse"
]);
-const MAP_OBJECT_PATHS = /* @__PURE__ */ new Set(["request_token_usage", "messages.Agent.tool_results"]);
+const MAP_OBJECT_PATHS = /* @__PURE__ */ new Set(["request_token_usage", "messages.Agent.tool_results", "acpx.session_options.env"]);
const OPAQUE_VALUE_PATHS = /* @__PURE__ */ new Set([
"agent_capabilities",
"messages.Agent.content.ToolUse.input",
@@ -2557,7 +2557,7 @@ function readCommandLineChar(state) {
escaping: false,
hasPart: true
};
- if (state.ch === "\\" && state.quote !== "'") return {
+ if (process.platform !== "win32" && state.ch === "\\" && state.quote !== "'") return {
current: state.current,
quote: state.quote,
escaping: true,
@@ -2912,8 +2912,8 @@ function promotePrefixedAuthEnvironment(env) {
}
return protectedKeys;
}
-function buildAgentEnvironment(authCredentials, sessionEnv) {
- const env = { ...process.env };
+function buildAgentEnvironment(authCredentials, sessionEnv, inheritProcessEnv = true) {
+ const env = inheritProcessEnv ? { ...process.env } : {};
const protectedAuthEnvKeys = promotePrefixedAuthEnvironment(env);
if (authCredentials) for (const [methodId, credential] of Object.entries(authCredentials)) {
addAuthCredentialEnvKeys(protectedAuthEnvKeys, methodId, credential);
@@ -2955,10 +2955,10 @@ function resolveConfiguredAuthCredential(methodId, authCredentials) {
const configCredentials = authCredentials ?? {};
return configCredentials[methodId] ?? configCredentials[toEnvToken(methodId)];
}
-function buildAgentSpawnOptions(cwd, authCredentials, sessionEnv) {
+function buildAgentSpawnOptions(cwd, authCredentials, sessionEnv, inheritProcessEnv) {
return {
cwd,
- env: buildAgentEnvironment(authCredentials, sessionEnv),
+ env: buildAgentEnvironment(authCredentials, sessionEnv, inheritProcessEnv),
stdio: [
"pipe",
"pipe",
@@ -3724,10 +3724,13 @@ function resolveClientCapabilities(params) {
},
terminal: params.terminal
};
- if (!params.devinAcp) return baseCapabilities;
+ const typedSessionFailureMeta = {
+ jetbrains: { air: { version: 1, capabilities: ["sessionFailure"] } }
+ };
+ if (!params.devinAcp) return { ...baseCapabilities, _meta: typedSessionFailureMeta };
return {
...baseCapabilities,
- _meta: DEVIN_COMPATIBILITY_CLIENT_CAPABILITIES_META
+ _meta: { ...typedSessionFailureMeta, ...DEVIN_COMPATIBILITY_CLIENT_CAPABILITIES_META }
};
}
function isDevinRequestDiagnosticsMethod(method) {
@@ -3957,9 +3960,26 @@ var AcpClient = class {
this.lastAgentExit = void 0;
this.lastKnownPid = child.pid ?? void 0;
this.attachAgentLifecycleObservers(child);
+ if (this.options.onAgentSpawn) {
+ try {
+ if (typeof child.pid !== "number" || child.pid <= 0) {
+ throw new Error("ACPX agent spawn did not expose a valid process id.");
+ }
+ await this.options.onAgentSpawn({ pid: child.pid, startedAt: this.agentStartedAt });
+ } catch (error) {
+ try {
+ child.kill("SIGKILL");
+ } catch {}
+ throw error;
+ }
+ }
const startupStderr = [];
child.stderr.on("data", (chunk) => {
this.captureStartupStderr(startupStderr, chunk);
+ if (this.options.onAgentStderr) {
+ this.options.onAgentStderr(chunk.toString());
+ return;
+ }
if (!this.options.verbose) return;
process.stderr.write(chunk);
});
@@ -3994,7 +4014,7 @@ var AcpClient = class {
geminiAcp: isGeminiAcpCommand(spawnCommand, args),
copilotAcp: isCopilotAcpCommand(spawnCommand, args),
claudeAcp: isClaudeAcpCommand(spawnCommand, args),
- spawnOptions: buildAgentSpawnOptions(this.options.cwd, this.options.authCredentials, this.options.sessionOptions?.env)
+ spawnOptions: buildAgentSpawnOptions(this.options.spawnCwd ?? this.options.cwd, this.options.authCredentials, this.options.sessionOptions?.env, this.options.inheritProcessEnv)
};
}
logAgentLaunch(plan) {
@@ -6070,6 +6090,27 @@ async function withConnectedSession(options) {
//#region src/runtime/engine/prompt-turn.ts
const SESSION_REPLY_IDLE_MS = 1e3;
const SESSION_REPLY_DRAIN_TIMEOUT_MS = 5e3;
+const TYPED_SESSION_FAILURE_CATEGORIES = /* @__PURE__ */ new Set([
+ "connection",
+ "access",
+ "limit",
+ "service",
+ "request",
+ "unknown"
+]);
+function typedTerminalSessionFailureCategory(response) {
+ if (response === null || typeof response !== "object" || Array.isArray(response)) return null;
+ const meta = response._meta;
+ if (meta === null || typeof meta !== "object" || Array.isArray(meta)) return null;
+ const jetbrains = meta.jetbrains;
+ if (jetbrains === null || typeof jetbrains !== "object" || Array.isArray(jetbrains)) return null;
+ const air = jetbrains.air;
+ if (air === null || typeof air !== "object" || Array.isArray(air)) return null;
+ if (!Number.isInteger(air.version) || air.version < 1) return null;
+ const failure = air.sessionFailure;
+ if (failure === null || typeof failure !== "object" || Array.isArray(failure) || failure.severity !== "error") return null;
+ return typeof failure.category === "string" && TYPED_SESSION_FAILURE_CATEGORIES.has(failure.category) ? failure.category : "unknown";
+}
async function runPromptTurn(params) {
try {
const promptPromise = params.client.prompt(params.sessionId, params.prompt);
@@ -6079,6 +6120,20 @@ async function runPromptTurn(params) {
idleMs: SESSION_REPLY_IDLE_MS,
timeoutMs: SESSION_REPLY_DRAIN_TIMEOUT_MS
}).catch(() => {});
+ const terminalFailureCategory = typedTerminalSessionFailureCategory(response);
+ if (terminalFailureCategory !== null) {
+ // Pass complete provider text to the diagnostic callback. Consumers redact
+ // before bounding it; truncating here can split and expose a credential.
+ const failure = response._meta.jetbrains.air.sessionFailure;
+ try {
+ params.onTerminalSessionFailure?.({
+ category: terminalFailureCategory,
+ ...(typeof failure.title === "string" ? { title: failure.title } : {}),
+ ...(typeof failure.details === "string" ? { details: failure.details } : {})
+ });
+ } catch {}
+ throw new Error(`ACP agent reported a terminal ${terminalFailureCategory} failure.`);
+ }
recordPromptResponseUsage(params.conversation, response.usage, params.promptMessageId);
return {
stopReason: response.stopReason,
diff --git a/dist/runtime.d.ts b/dist/runtime.d.ts
index ccdbe5b032521518022223733049b8b38793473b..416bdaae1450c6e85b34afd154349770f9f7b0a9 100644
--- a/dist/runtime.d.ts
+++ b/dist/runtime.d.ts
@@ -45,6 +45,7 @@ type AcpRuntimeTurnAttachment = {
data: string;
};
type AcpRuntimeTurnInput = {
+ onTerminalSessionFailure?: (failure: { category: string; title?: string; details?: string }) => void;
handle: AcpRuntimeHandle;
text: string;
attachments?: AcpRuntimeTurnAttachment[];
@@ -266,6 +267,10 @@ type AcpRuntimeOptions = {
timeoutMs?: number;
probeAgent?: string;
verbose?: boolean;
+ onAgentStderr?: (chunk: string) => void;
+ onAgentSpawn?: (meta: { pid: number; startedAt: string }) => Promise<void>;
+ spawnCwd?: string;
+ inheritProcessEnv?: boolean;
onPermissionRequest?: (req: AcpPermissionRequest, ctx: {
signal: AbortSignal;
}) => Promise<AcpPermissionDecision | undefined>;
@@ -302,6 +307,7 @@ declare class AcpRuntimeManager {
private createAndSaveRuntimeRecord;
private keepPersistentClient;
startTurn(input: {
+ onTerminalSessionFailure?: (failure: { category: string; title?: string; details?: string }) => void;
handle: AcpRuntimeHandle;
text: string;
attachments?: AcpRuntimeTurnAttachment[];
@@ -334,6 +340,7 @@ declare class AcpRuntimeManager {
private finalizeRuntimeTurn;
private finalizeRuntimeTurnRecord;
runTurn(input: {
+ onTerminalSessionFailure?: (failure: { category: string; title?: string; details?: string }) => void;
handle: AcpRuntimeHandle;
text: string;
attachments?: AcpRuntimeTurnAttachment[];
diff --git a/dist/runtime.js b/dist/runtime.js
index 6c9cc999e50a11c399c68b3a0f1b7af4bc2317c0..e06bf4fdee534f3ef56be0387a2f5f8626e697ad 100644
--- a/dist/runtime.js
+++ b/dist/runtime.js
@@ -744,7 +744,8 @@ var AcpRuntimeManager = class {
this.deps = deps;
}
createClient(options) {
- return this.deps.clientFactory?.(options) ?? new AcpClient(options);
+ const clientOptions = { ...options, onAgentStderr: this.options.onAgentStderr, onAgentSpawn: this.options.onAgentSpawn, spawnCwd: this.options.spawnCwd, inheritProcessEnv: this.options.inheritProcessEnv };
+ return this.deps.clientFactory?.(clientOptions) ?? new AcpClient(clientOptions);
}
async readPendingPersistentClient(record, options) {
const pendingClient = this.pendingPersistentClients.get(record.acpxRecordId);
@@ -982,6 +983,7 @@ var AcpRuntimeManager = class {
client: turn.client,
sessionId,
prompt: task.promptInput,
+ onTerminalSessionFailure: task.input.onTerminalSessionFailure,
timeoutMs: task.input.timeoutMs ?? this.options.timeoutMs,
conversation: turn.conversation,
promptMessageId: turn.promptMessageId
@@ -1586,6 +1588,7 @@ var AcpxRuntime = class {
const turnPromise = this.getManager().then((manager) => manager.startTurn({
handle,
text: input.text,
+ onTerminalSessionFailure: input.onTerminalSessionFailure,
attachments: input.attachments,
mode: input.mode,
sessionMode: state.mode,
@@ -1614,6 +1617,7 @@ var AcpxRuntime = class {
yield* (await this.getManager()).runTurn({
handle,
text: input.text,
+ onTerminalSessionFailure: input.onTerminalSessionFailure,
attachments: input.attachments,
mode: input.mode,
sessionMode: state.mode,
diff --git a/dist/session-options-jkYbBxGE.d.ts b/dist/session-options-jkYbBxGE.d.ts
index 9d37f377fb6a0828e0d2bc5a48754f3aa71509a4..22dda59e0b7616a20ad45218913e5d8f884bae49 100644
--- a/dist/session-options-jkYbBxGE.d.ts
+++ b/dist/session-options-jkYbBxGE.d.ts
@@ -84,6 +84,10 @@ type AcpClientOptions = {
terminal?: boolean;
suppressSdkConsoleErrors?: boolean;
verbose?: boolean;
+ onAgentStderr?: (chunk: string) => void;
+ onAgentSpawn?: (meta: { pid: number; startedAt: string }) => Promise<void>;
+ spawnCwd?: string;
+ inheritProcessEnv?: boolean;
sessionOptions?: {
model?: string;
allowedTools?: string[];