Active goal turns that stopped with text could hit the empty/unexpected-stop
continuation guards before threshold maintenance. When those guards scheduled
another goal turn, #checkCompaction never ran, so no auto_compaction_start was
emitted even while visible context stayed above thresholdTokens.
Run threshold maintenance once before those active-goal self-continuations and
log the threshold decision inputs: billed context, stored estimate, resolved
trigger tokens, post-maintenance tokens, strategy, threshold, promotion state,
and shouldCompact.
Also pass post-prune maintenance tokens into the shake recovery-band check so
supersede/drop-useless savings are preserved when deciding whether shake still
needs to fall back to context-full compaction.
Fixes#3174
Pruning frees bytes for the NEXT prompt — it does not change the size of
the prompt the LLM just billed for. Subtracting the per-turn
`#pruneStaleToolResults` / `#pruneToolOutputs` savings from the
threshold input let a long-running `/goal` session sit above
`compaction.thresholdTokens` indefinitely: the visible context
(anchored to the same provider billing) showed >threshold, but
`shouldCompact` no-op'd because the subtraction dropped the input below
the trigger. The `compactionContextTokens` floor against the post-prune
local estimate is still applied, so a payload-compression hook still
can't deflate the trigger.
Regression test seeds one large `useless` tool result whose suffix sits
inside the 8k cache-warm window so `#pruneStaleToolResults` actually
returns ≥20k savings, then asserts compaction fires when the final turn
bills 91k tokens against the reporter's `thresholdTokens: 76384`.
Fixes#3174
- Replaced the `/debug dump-next-request` command with an updated `/dump` command that exports LLM request context to JSON sidecar files.
- Removed persistent debug path state and manual path configuration in favor of automated generation.
- Updated session logic to handle serializing LLM request context to temporary directories.
- Refactored testing suites to remove path-based debug tests and verify dynamic request file generation.
- Updated `render`, `renderMany`, and native snapcompact methods to return promises, ensuring scalable async execution.
- Refactored `transformProviderContext` and `buildSideRequestContext` to support asynchronous operations in agent loops.
- Integrated `Promise.all` for improved concurrency when processing frame rendering and rendering batch operations.
- Updated all internal call sites, SDK hooks, and test suites to accommodate the asynchronous API signatures.
- Added validation to scan for non-ASCII characters before performing snap-compaction, falling back to LLM-based summarization if the unrenderable ratio is too high.
- Updated event handling and status reporting to explicitly support snapcompact actions, including specific error warnings and cancellation states in the UI.
- Updated session logic to default to snapcompact strategy when auto-compaction is enabled.
CI caught that flooring by the raw local estimate falsely triggers compaction on
thinking-heavy turns: estimateTokens counts the opaque thinkingSignature /
redactedThinking payloads (providers bill them on replay, #2275), but their local
byte size diverges wildly from what the provider actually charges — so a turn
with a large encrypted-reasoning blob but small provider usage would trip the
floor (broke agent-session-handoff 'provider-anchored usage' test).
estimateTokens now takes { excludeEncryptedReasoning } and the compaction floor
(#estimateStoredContextTokens) uses it: the floor counts only reliably-countable,
on-wire-compressible content (text, tool results, tool calls), while the provider
usage arm of compactionContextTokens still accounts for encrypted reasoning. This
keeps the encrypted-reasoning case provider-anchored while still flooring upward
when a before_provider_request hook compresses tool results.
A before_provider_request extension (a context-compression proxy like Headroom,
an obfuscator, or inline snapcompact) can shrink the outgoing request below the
real stored conversation. The provider then reports deflated prompt tokens, so
the auto-compaction threshold never fires and the stored history grows unbounded
until it overflows the context window and can no longer be compacted at all.
Add compactionContextTokens(provider, storedEstimate) = max(provider, estimate)
and apply it to both the pre-prompt and post-response compaction decisions,
flooring the provider-reported tokens by the agent's own estimate of the stored
conversation. Display and cost accounting still use exact provider usage; only
the compaction trigger takes the floor.
- Added `buildSideRequestContext` to the `Agent` class to generate prompt-cache-friendly provider contexts.
- Updated ephemeral side-channel turns to forward the full tool catalog to maintain prompt cache hit rates.
- Injected a `developer` role reminder into ephemeral turns to instruct the model to suppress tool calls.
- Implemented automatic post-processing to strip any tool calls from ephemeral turn responses.
- Exported message and dialect helper functions in `agent-loop.ts` to support context construction.
- Added a set to track tool result IDs that have been rewound to ensure they are not re-appended to the session history.
- Updated message handling logic to conditionally skip persistence for tool results associated with active rewind operations.
- Implemented message deduplication in `AdvisorRuntime` to collapse verbatim re-injected primary context (plan rules/approved plans) between turns.
- Reduced token consumption by replacing identical primary context segments with a status marker.
- Updated history formatter to allow expansion of load-bearing context types specifically, while retaining one-line summaries for other custom messages.
- Fixed session history desynchronization during `rewind` by performing a full context rebuild after applying branch changes.
- Prevented `rewind` tool output and assistant side-channel data from polluting the prompt cache by flushing and sanitizing session state.
- Added explicit test coverage for context reconstruction and assistant message sanitization after rewind events.
fork() reset mnemopi conversation tracking directly but skipped the shared new-transcript reset, so the folded/promoted first-turn memory stayed in #baseSystemPrompt. The next turn re-recalled and the change-detection saw no diff, taking the fallback promotion path and injecting the <memories> block twice into the forked session prompt. Route fork() through #resetMemoryContextForNewTranscript() like the other reset paths and add a regression test asserting the forked prompt contains recalled memory exactly once.
Run threshold compaction maintenance when an active goal turn ends through a successful yield, while preserving the final-yield skip for non-goal completions.
Fixes#3146
Restored and refreshed promoted memory prompts through the shared new-transcript reset path used by new sessions, handoff, branch, /btw, and cross-session switches.\n\nAdded coverage for newSession after first-turn memory recall so stale recalled memories cannot leak into the next transcript when recall returns no context.\n\nFixes #3111
Reset memory recall state before rebuilding the prompt during session switches, and clear fallback-promoted memory prompts when moving to another session.\n\nAdded a regression test that switches sessions after first-turn memory recall and verifies the next session does not receive stale memories.\n\nFixes #3111
Promoted first-turn memory recall into the stable base prompt so append-only sessions do not drop the memory block on the next turn and rebuild the provider prefix.\n\nAdded a regression test covering a memory backend that recalls once before the first model request.\n\nFixes #3111
- Updated `buildOpenAICompat` to override the `qwen` thinking format for Fireworks-hosted models, ensuring they use `openai` thinking parameters instead.
- Prevented invalid `enable_thinking` payload errors by ensuring Fireworks-hosted Qwen requests conform to their strict schema.
- Updated `AgentSession` to allow Fireworks fast-fallback logic to execute even when standard retries are disabled.
- Added support for "Fast" serving-path variants for select Fireworks models.
- Updated compatibility logic to route `-fast` suffixes to the appropriate router wire format.
- Extended the model generation catalog to include these Fast variants with their respective pricing.
- Updated AI types to allow the `priority` service tier for Fireworks providers.
- AsyncDrain reset #queue before invoking the handler, risking lost pushes.
- Moved queue reset to after the handler runs and wrapped exec in a promise.
- Added history-storage drain tests covering batch flush and fresh-batch behavior.
- Privatized the legacy `nextToolChoice` method to `#nextHardToolChoice` to ensure all tool-choice directives flow through the unified `nextToolChoiceDirective` entry point.
- Eliminated redundant dual entry points for fetching tool choices, which previously bypassed the soft pending-preview lifecycle.
- Updated test suites to consume `nextToolChoiceDirective` where appropriate to maintain consistency with internal agent-loop logic.
Cleared pending-preview markers when resolve has no runnable handler so a stale gate cannot keep forcing resolve after the invoker is gone.
Added regression coverage for apply and discard draining stale pending markers.
Fixes#3061
- Improved thinking loop detection logic by refining text normalization and treating stalls as retryable errors.
- Instrumented the agent session to recognize thinking loop markers within retryable error conditions.
- Automated the clearing of stale error banners upon successful auto-retry execution.
- Added comprehensive test coverage for chunked thinking loop errors and banner management.
- Replaced usage of `ReturnType<typeof setTimeout>` and `ReturnType<typeof setInterval>` with the explicit `Timer` type across the codebase.
- Updated several type definitions and function signatures to use concrete types instead of inferred return types for improved clarity and maintainability.
- Implement JSON repair and strict argument validation to sanitize raw payloads and redact sensitive information from agent event logs.
- Add automatic authentication fallback for benchmark model resolution to ensure consistent performance testing across providers.
- Refactor search tool API parameters by replacing `i` with a case-sensitive `case` boolean flag for clarity.
- Update session history formatting to ensure empty objects are consistently serialized as `{}` instead of empty strings.
- Added an early return to stop event propagation if no handlers are registered for the event type.
- Updated the `agent_start` logic to return early, ensuring consistent flow control.
- Introduced a centralized `discoverAuthStorage` mechanism across packages to unify credential retrieval and configuration resolution.
- Added support for new Gemini and Moonshot model variants while updating context window and effort configuration for existing models.
- Resolved provider-specific 400 errors for OpenRouter and GLM models by refining reasoning effort mapping and retry logic.
- Standardized credential management in both the coding-agent and model catalog by migrating to the unified authentication broker.
- Export `directoryExists` in `utils` to safely validate working directories before traversal.
- Update `SessionManager` and startup logic to fallback to the launch directory if a session's recorded working directory no longer exists.
- Add regression tests to ensure sessions now correctly adopt the launch directory instead of crashing on missing paths.
- Add `directoryExists` check to validate session working directories.
- Skip updating the session cwd if the recorded directory no longer exists on disk to prevent runtime errors.
Moved mnemopi's local embedding provider out of the main agent process by
spawning a dedicated `Bun.spawn` child for fastembed + onnxruntime-node. The
agent CLI gains a hidden `__omp_worker_mnemopi_embed` dispatch; `loadMnemopi`
/ `loadMnemopiCore` install the subprocess-backed initializer through the
newly-exposed `setLocalModelInitializer` seam so every `embed()` call
round-trips through IPC instead of loading the NAPI module. The parent
SIGKILLs the child on dispose so the destructor that segfaults Bun on
Windows shutdown (NAPI finalizer at exit on npm installs, `process.dlopen`
constructor at session start on standalone binaries) never runs in any
address space the agent owns. Mirrors the tiny-model fix from #1607.
Adds `smokeTestMnemopiEmbedWorker` to `omp --smoke-test`, a new
`test/issue-3031-repro.test.ts` that pins the spawn/dispatch/signal-exit
contract and forbids re-importing `fastembed-runtime` from the agent surface,
and changelog entries.
Fixes#3031
- Replaced four individual power boolean settings with a single `power.sleepPrevention` enum for improved configuration management.
- Implemented automatic migration logic in `Settings.init` to map legacy macOS power booleans to the new enum levels.
- Updated `AgentSession` to utilize the new `sleepPrevention` enum for macOS power assertion lifecycle management.
- Changed default value of `display.cacheMissMarker` to `false` to suppress cache-miss markers.
- Added comprehensive unit tests in `settings-manager.test.ts` to verify backward-compatible migration behavior.
- Added tracking for expected cache invalidations during model changes, compactions, and plan-mode transitions.
- Included `cacheMissExplainedAt` metadata in session context to prevent displaying misleading cache miss warnings in the transcript.
- Updated controller logic to reset assistant usage markers when mode-switching or performing actions that invalidate the prompt cache.
- Renamed `repeatToolDescriptions` to `inlineToolDescriptors` throughout configuration, SDK, and internal session management.
- Set the default value to `true` and updated descriptions to clarify the descriptor inlining behavior.
- Fixed the `/dump` command to prevent duplicate tool inventory output when inlining is enabled.
- Introduced `SoftToolRequirement` to support non-invasive tool enforcement with lifecycle management and escalation.
- Added `ToolChoiceDirective` to coordinate hard and soft tool requirements within the agent loop.
- Optimized preview workflows in `coding-agent` by replacing forced tool choices with non-forcing pending invokers.
- Enhanced `CompactionSummaryMessage` to prioritize structured rendering for tool requirement reminders.
- Changed `wrapSteeringForModel` to wrap every `steering:true` user message regardless of position, so a steer's wire bytes stay identical once buried instead of reverting to raw and busting the prompt cache.
- Neutralized the present-tense framing in `user-interjection.md` so always-wrapping no longer leaves stale "current task" wording on buried interjections.
- Updated `session-messages.test.ts` to assert buried steering messages are wrapped too.
- Added `keepBoundaryId` and `cacheWarmSuffixTokens` guards to `pruneToolOutputs` and `pruneSupersededToolResults` so superseded/useless results sitting in the already-sent cached prefix are no longer rewritten mid-session, with new `computeMessageSuffixTokens`/`resolveBoundaryIndex` helpers.
- Added a `keepBoundaryId` floor to `collectShakeRegions` so shake skips entries summarized away by the latest compaction.
- Threaded `firstKeptEntryId`, `PRUNE_CACHE_WARM_SUFFIX_TOKENS` (8k) and `PRUNE_IDLE_FLUSH_MS` (90m, above the 1h cache TTL) from `#pruneToolOutputs`, `#pruneStaleToolResults` and `shake` in `agent-session.ts`.
- Added six boundary tests in `supersede-prune.test.ts` covering warm-prefix protection, tail-case pruning, and the pre-boundary floor.
- Added utility functions to strip descriptions from JSON schemas and tool definitions for optimized token output.
- Integrated `pruneToolDescriptions` configuration across agent loops and sessions to enable optional schema pruning.
- Updated agent context and snapshot logic to propagate pruning settings and maintain fingerprinting integrity.
- Verified schema structural integrity and removal of annotation descriptions through new unit tests.
- Update `MAX_FRAMES_DEFAULT` to 80 to better utilize high-capacity model context windows.
- Remove `providerFrameBudget` from `snapcompact` to decouple archival limits from provider-specific image caps.
- Update `compact` logic to treat `maxFrames` as a hard upper bound rather than a provider-clamped limit.
Fix all Windows-specific test failures caused by path handling problems
and EBUSY errors from unclosed SQLite database handles.
Root causes fixed:
1. POSIX path assumptions: replaced hard-coded file:///tmp, /repo, etc.
with pathToFileURL/path.resolve/path.join computed expectations
2. shortenPath() now normalizes backslashes to forward slashes after ~
and respects home directory boundaries
3. HistoryStorage.resetInstance() leaked its Database — added #close()
that finalizes all prepared statements and closes the DB
4. AgentStorage gained the same resetInstance()/#close() pattern
5. SqliteAuthCredentialStore.close() leaked one-off prepared statements
from inline this.#db.prepare() calls — wrapped each in try/finally
6. model-cache.ts used a process-global DB even for custom dbPath —
now opens/closes per-call via withModelCacheDb
7. createAgentSession leaked AuthStorage on construction failure —
added ownsAuthStorage cleanup in catch block
8. MnemopiBackend.removeDbFiles() now truly best-effort (catches errors)
9. TempDir retry window expanded from 4x10ms to 40x25ms
10. TempDir prefix convention: non-@ prefixes created dirs relative to
cwd instead of os.tmpdir() — all test temp dirs now use @ prefix
11. Shell-escaped interpolated paths in bash tool tests
12. git core.autocrlf false in autoresearch test repo init
All 522 previously-failing Windows tests now pass.
- Moved the `INTENT_FIELD` constant from `@oh-my-pi/pi-agent-core` to the specialized `@oh-my-pi/pi-wire` package to permit broader usage across the monorepo.
- Updated all references across `agent`, `ai`, `coding-agent`, `collab-web`, and `snapcompact` packages to import the constant from the new location.
- Added `@oh-my-pi/pi-wire` as a dependency to all affected packages.
- Standardized elision markers across all tool outputs and filters to use cohesive `[...N [type] elided...]`, `[...Nln elided...]`, and `[...xB elided...]` syntax.
- Updated documentation, prompts, and test expectations to reflect the unified elision format.
- Improved transcript viewer robustness by preventing content aliasing through path-inclusive signature hashing.
- Added logic to clear stale transcript content when associated session files are deleted, accompanied by verifying test cases.
- Implemented `AdvisorTranscriptRecorder` to persist advisor sessions to append-only `__advisor.jsonl` files.
- Integrated transcript recording into agent sessions with managed flushing, atomic file switching, and synthetic turn attribution.
- Restricted advisor-kind agents by excluding them from rosters, history protocols, messaging, and interactive agent commands.
- Reserved the `__advisor` filename stem across the output manager and task registry to prevent task ID collisions.
roboomp