* initial commit
* feat: implement deriveSessionTotalTokens function and update usage tests
* Added deriveSessionTotalTokens function to calculate total tokens based on usage and context tokens.
* Updated usage tests to include cases for derived session total tokens.
* Refactored session usage calculations in multiple files to utilize the new function for improved accuracy.
* fix: restore overflow truncation fallback + changelog/test hardening (#11551) (thanks @tyler6204)
* fix(cron): comprehensive cron scheduling and delivery fixes
- Fix delivery target resolution for isolated agent cron jobs
- Improve schedule parsing and validation
- Add job retry logic and error handling
- Enhance cron ops with better state management
- Add timer improvements for more reliable cron execution
- Add cron event type to protocol schema
- Support cron events in heartbeat runner (skip empty-heartbeat check,
use dedicated CRON_EVENT_PROMPT for relay)
* fix: remove cron debug test and add changelog/docs notes (#11641) (thanks @tyler6204)
* fix: use STATE_DIR instead of hardcoded ~/.openclaw for identity and canvas
device-identity.ts and canvas-host/server.ts used hardcoded
path.join(os.homedir(), '.openclaw', ...) ignoring OPENCLAW_STATE_DIR
env var and the resolveStateDir() logic from config/paths.ts.
This caused ~/.openclaw/identity and ~/.openclaw/canvas directories
to be created even when state dir was overridden or resided elsewhere.
* fix: format and remove duplicate imports
* fix: scope state-dir patch + add regression tests (#4824) (thanks @kossoy)
* fix: align state-dir fallbacks in hooks and agent paths (#4824) (thanks @kossoy)
---------
Co-authored-by: Gustavo Madeira Santana <gumadeiras@gmail.com>
* fix: gracefully handle oversized tool results causing context overflow
When a subagent reads a very large file or gets a huge tool result (e.g.,
gh pr diff on a massive PR), it can exceed the model's context window in
a single prompt. Auto-compaction can't help because there's no older
history to compact — just one giant tool result.
This adds two layers of defense:
1. Pre-emptive: Hard cap on tool result size (400K chars ≈ 100K tokens)
applied in the session tool result guard before persistence. This
prevents extremely large tool results from being stored in full,
regardless of model context window size.
2. Recovery: When context overflow is detected and compaction fails,
scan session messages for oversized tool results relative to the
model's actual context window (30% max share). If found, truncate
them in the session via branching (creating a new branch with
truncated content) and retry the prompt.
The truncation preserves the beginning of the content (most useful for
understanding what was read) and appends a notice explaining the
truncation and suggesting offset/limit parameters for targeted reads.
Includes comprehensive tests for:
- Text truncation with newline-boundary awareness
- Context-window-proportional size calculation
- In-memory message truncation
- Oversized detection heuristics
- Guard-level size capping during persistence
* fix: prep fixes for tool result truncation PR (#11579) (thanks @tyler6204)
- Added tests to ensure proper sanitization of context overflow errors.
- Introduced a new function to determine when to rewrite context overflow messages.
- Updated the sanitization logic to improve user experience by providing clearer error messages while preserving conversational context.
* refactor: update cron job wake mode and run mode handling
- Changed default wake mode from 'next-heartbeat' to 'now' in CronJobEditor and related CLI commands.
- Updated cron-tool tests to reflect changes in run mode, introducing 'due' and 'force' options.
- Enhanced cron-tool logic to handle new run modes and ensure compatibility with existing job structures.
- Added new tests for delivery plan consistency and job execution behavior under various conditions.
- Improved normalization functions to handle wake mode and session target casing.
This refactor aims to streamline cron job configurations and enhance the overall user experience with clearer defaults and improved functionality.
* test: enhance cron job functionality and UI
- Added tests to ensure the isolated agent correctly announces the final payload text when delivering messages via Telegram.
- Implemented a new function to pick the last deliverable payload from a list of delivery payloads.
- Enhanced the cron service to maintain legacy "every" jobs while minute cron jobs recompute schedules.
- Updated the cron store migration tests to verify the addition of anchorMs to legacy every schedules.
- Improved the UI for displaying cron job details, including job state and delivery information, with new styles and layout adjustments.
These changes aim to improve the reliability and user experience of the cron job system.
* test: enhance sessions thinking level handling
- Added tests to verify that the correct thinking levels are applied during session spawning.
- Updated the sessions-spawn-tool to include a new parameter for overriding thinking levels.
- Enhanced the UI to support additional thinking levels, including "xhigh" and "full", and improved the handling of current options in dropdowns.
These changes aim to improve the flexibility and accuracy of thinking level configurations in session management.
* feat: enhance session management and cron job functionality
- Introduced passthrough arguments in the test-parallel script to allow for flexible command-line options.
- Updated session handling to hide cron run alias session keys from the sessions list, improving clarity.
- Enhanced the cron service to accurately record job start times and durations, ensuring better tracking of job execution.
- Added tests to verify the correct behavior of the cron service under various conditions, including zero-delay timers.
These changes aim to improve the usability and reliability of session and cron job management.
* feat: implement job running state checks in cron service
- Added functionality to prevent manual job runs if a job is already in progress, enhancing job management.
- Updated the `isJobDue` function to include checks for running jobs, ensuring accurate scheduling.
- Enhanced the `run` function to return a specific reason when a job is already running.
- Introduced a new test case to verify the behavior of forced manual runs during active job execution.
These changes aim to improve the reliability and clarity of cron job execution and management.
* feat: add session ID and key to CronRunLogEntry model
- Introduced `sessionid` and `sessionkey` properties to the `CronRunLogEntry` struct for enhanced tracking of session-related information.
- Updated the initializer and Codable conformance to accommodate the new properties, ensuring proper serialization and deserialization.
These changes aim to improve the granularity of logging and session management within the cron job system.
* fix: improve session display name resolution
- Updated the `resolveSessionDisplayName` function to ensure that both label and displayName are trimmed and default to an empty string if not present.
- Enhanced the logic to prevent returning the key if it matches the label or displayName, improving clarity in session naming.
These changes aim to enhance the accuracy and usability of session display names in the UI.
* perf: skip cron store persist when idle timer tick produces no changes
recomputeNextRuns now returns a boolean indicating whether any job
state was mutated. The idle path in onTimer only persists when the
return value is true, eliminating unnecessary file writes every 60s
for far-future or idle schedules.
* fix: prep for merge - explicit delivery mode migration, docs + changelog (#10776) (thanks @tyler6204)
* feat(antigravity): update default model to Claude Opus 4.6
Claude Opus 4.5 has been replaced by Claude Opus 4.6 on the
Antigravity (Google Cloud Code Assist) platform.
- Update DEFAULT_MODEL in google-antigravity-auth extension
- Update testing docs to reference the new model
* fix: update remaining antigravity opus 4.5 refs in zh-CN docs and tests
Address review comments from Greptile:
- Update zh-CN/testing.md antigravity model references
- Update pi-tools-agent-config.test.ts model IDs
* Antigravity: default OAuth model to Opus 4.6 (#10720) (thanks @calvin-hpnet)
---------
Co-authored-by: Tak Hoffman <781889+Takhoffman@users.noreply.github.com>
* feat(memory): add native Voyage AI embedding support with batching
Cherry-picked from PR #2519, resolved conflict in memory-search.ts
(hasRemote -> hasRemoteConfig rename + added voyage provider)
* fix(memory): optimize voyage batch memory usage with streaming and deduplicate code
Cherry-picked from PR #2519. Fixed lint error: changed this.runWithConcurrency
to use imported runWithConcurrency function after extraction to internal.ts
* fix(docs): correct OpenCode Zen description in code comment
OpenCode Zen is a pay-as-you-go token-based API, not a $200/month
subscription. The subscription tiers ($20/$100/$200) are OpenCode Black,
a separate product.
This fixes the misleading comment that conflated Zen with Black.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
* fix: align OpenCode Zen billing copy (#9998) (thanks @therealZpoint-bot)
---------
Co-authored-by: Claude <claude@archibald.local>
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
Co-authored-by: Gustavo Madeira Santana <gumadeiras@gmail.com>
* fix(ollama): add streaming config and fix OLLAMA_API_KEY env var support
Adds configurable streaming parameter to model configuration and sets streaming
to false by default for Ollama models. This addresses the corrupted response
issue caused by upstream SDK bug badlogic/pi-mono#1205 where interleaved
content/reasoning deltas in streaming responses cause garbled output.
Changes:
- Add streaming param to AgentModelEntryConfig type
- Set streaming: false default for Ollama models
- Add OLLAMA_API_KEY to envMap (was missing, preventing env var auth)
- Document streaming configuration in Ollama provider docs
- Add tests for Ollama model configuration
Users can now configure streaming per-model and Ollama authentication
via OLLAMA_API_KEY environment variable works correctly.
Fixes#8839
Related: badlogic/pi-mono#1205
* docs(ollama): use gpt-oss:20b as primary example
Updates documentation to use gpt-oss:20b as the primary example model
since it supports tool calling. The model examples now show:
- gpt-oss:20b as the primary recommended model (tool-capable)
- llama3.3 and qwen2.5-coder:32b as additional options
This provides users with a clear, working example that supports
OpenClaw's tool calling features.
* chore: remove unused vi import from ollama test
Adds cleanup handlers to release held file locks when the process
terminates via SIGTERM, SIGINT, or normal exit. This prevents orphaned
lock files that would block future sessions.
Fixes#1951
Fixes tool call/tool_result pairing issues that cause permanent session corruption when assistant messages have stopReason "error" or "aborted". Includes 4 unit tests.
* fix(errors): return clear billing error message instead of cryptic raw error (#8136)
When an LLM API provider returns a credit/billing-related error (HTTP 402,
insufficient credits, low balance, etc.), OpenClaw now shows a clear,
actionable message instead of passing through the raw/cryptic error text:
⚠️ API provider returned a billing error — your API key has run out of
credits or has an insufficient balance. Check your provider's billing
dashboard and top up or switch to a different API key.
Changes:
- formatAssistantErrorText: detect billing errors via isBillingErrorMessage()
and return a user-friendly message (placed before the generic HTTP/JSON
error fallthrough)
- sanitizeUserFacingText: same billing detection for the sanitization path
- pi-embedded-runner/run.ts: add billingFailure detection in the profile
exhaustion fallback, so the FailoverError message is billing-specific
- Added 3 new tests for credit balance, HTTP 402, and insufficient credits
* fix: extract billing error message to shared constant
Previously, overflowCompactionAttempted was a boolean flag set once, preventing
recovery when a single compaction wasn't enough. Change to a counter allowing up
to 3 attempts before giving up. Also add diagnostic logging on overflow events to
help debug early-overflow issues.
Fixes sessions that hit context overflow during long agentic turns with many tool
calls, where one compaction round isn't sufficient to bring context below limits.
* fix(cron): prevent timer from allowing process exit (fixes#9694)
The cron timer was using .unref(), which caused the Node.js event
loop to exit or sleep if no other handles were active. This prevented
cron jobs from firing in some environments.
* fix(cron): infer delivery target for isolated jobs (fixes#9683)
When creating isolated agentTurn jobs (e.g. reminders) without explicit
delivery options, the job would default to 'announce' but fail to
resolve the target conversation. Now, we infer the channel and
recipient from the agent's current session key.
* fix(cron): enhance delivery inference for threaded sessions and null inputs (#9733)
Improves the delivery inference logic in the cron tool to correctly handle threaded session keys and cases where delivery is explicitly set to null. This ensures that the appropriate delivery mode and target are inferred based on the agent's session key, enhancing the reliability of job execution.
* fix: preserve telegram topic delivery inference (#9733) (thanks @tyler6204)
* fix: simplify cron delivery merge spread (#9733) (thanks @tyler6204)
When pruneHistoryForContextShare drops chunks of messages, it could drop
an assistant message with tool_use blocks while leaving corresponding
tool_result messages in the kept portion. These orphaned tool_results
cause Anthropic's API to reject the session with 'unexpected tool_use_id'.
Fix by calling repairToolUseResultPairing after each chunk drop to clean
up any orphaned tool_results. This reuses existing battle-tested code
from session-transcript-repair.ts.
Fixes#9769, #9724, #9672
* feat: add Claude Opus 4.6 to built-in model catalog
- Update default model from claude-opus-4-5 to claude-opus-4-6
- Add opus-4.6 model ID normalization
- Add claude-opus-4-6 to live model filter prefixes
- Update image tool to prefer claude-opus-4-6 for vision
- Add CLI backend alias for opus-4.6
- Update onboard auth default selections to include opus-4.6
- Update model picker placeholder
Closes#9811
* test: update tests for claude-opus-4-6 default
- Fix model-alias-defaults test to use claude-opus-4-6
- Fix image-tool test to expect claude-opus-4-6 in fallbacks
* feat: support claude-opus-4-6
* docs: update changelog for opus 4.6 (#9853) (thanks @TinyTb)
* chore: bump pi to 0.52.0
---------
Co-authored-by: Slurpy <slurpy@openclaw.ai>
Co-authored-by: Peter Steinberger <steipete@gmail.com>
When spawning a subagent, the requesterOrigin's threadId, to, and
accountId were not forwarded to the callGateway({method:'agent'}) params.
This meant the subagent's runContext had no currentThreadTs or
currentChannelId, so resolveTelegramAutoThreadId could not auto-inject
the forum topic thread ID when the subagent used the message tool.
Changes:
- sessions-spawn-tool: pass to, accountId, threadId from requesterOrigin
- run-context: populate currentChannelId from opts.to as fallback
Fixes subagent messages landing in General Topic instead of the correct
Telegram DM topic thread.
* Message: enforce sandbox for media param
* fix: harden sandboxed media handling (#8780) (thanks @victormier)
* chore: format message action runner (#8780) (thanks @victormier)
---------
Co-authored-by: Victor Mier <victormier@gmail.com>
- Introduced new properties for explicit message targeting and message tool disabling in the EmbeddedRunAttemptParams type.
- Updated cron job tests to validate best-effort delivery behavior and handling of delivery failures.
- Added logic to clear delivery settings when switching session targets in cron jobs.
- Improved the resolution of delivery failures and best-effort logic in the isolated agent's run function.
This update enhances the flexibility and reliability of delivery mechanisms in isolated cron jobs, ensuring better handling of message delivery scenarios.
- Updated isolated cron jobs to support new delivery modes: `announce` and `none`, improving output management.
- Refactored job configuration to remove legacy fields and streamline delivery settings.
- Enhanced the `CronJobEditor` UI to reflect changes in delivery options, including a new segmented control for delivery mode selection.
- Updated documentation to clarify the new delivery configurations and their implications for job execution.
- Improved tests to validate the new delivery behavior and ensure backward compatibility with legacy settings.
This update provides users with greater flexibility in managing how isolated jobs deliver their outputs, enhancing overall usability and clarity in job configurations.
- Default one-shot jobs to delete after success, improving job management.
- Introduced `--keep-after-run` CLI option to allow users to retain one-shot jobs post-execution.
- Updated documentation to clarify default behaviors and new options for one-shot jobs.
- Adjusted cron job creation logic to ensure consistent handling of delete options.
- Enhanced tests to validate new behaviors and ensure reliability.
This update streamlines the handling of one-shot jobs, providing users with more control over job persistence and execution outcomes.
