- Added `loadOverallPlanReference` to resolve a session plan reference from local storage and skip empty or missing files.
- Updated task execution to read the active plan reference (except in plan mode) and pass it into each spawned subagent.
- Extended the subagent system prompt and session SDK/tools plumbing so subagents receive and render the approved plan path and contents.
- Added `isReadOnlyAgent` and `READ_ONLY_TOOL_NAMES` to classify agents.
- Marked read-only agents and forbade edits, commands, and reasoning offload.
- Added tests for capability classification and description rendering.
- Added `repairDoubleEncodedJsonString` to unescape fields double-encoded by the model (e.g. literal `\n`, `\"`, `\uXXXX` in `context`/`assignment`/`description`).
- Scoped repair to natural-language fields only, leaving code-bearing tools untouched.
- Applied repair on both render and execution paths in `TaskTool`.
- Stored mcpManager and localProtocolOptions on ToolSession so nested subagents inherit them without relying on process-global singletons.
- TaskTool now uses the session's localProtocolOptions and mcpManager when spawning sub-tasks, falling back to defaults if absent.
- Added parentMnemosyneSessionState propagation from session state through SDK, executor, and task options into nested sessions.
- Added getMnemosyneSessionState() and rekeying logic to refresh Mnemosyne IDs during session sync, switch, and restore.
- Added Mnemosyne reset and teardown cleanup on unaliasing or restoration to avoid stale state.
- Added `ToolTier`, `ToolApproval`, and `ToolApprovalDecision` types and exported approval APIs.
- Updated approval-mode options from `auto|prompt|custom` to `always-ask|write|yolo` and defaulted mode to `yolo`.
- Changed approval resolution to apply per-tool decisions first, then mode-tier limits, with legacy-mode migration.
- Assigned read/write/exec `approval` and approval-detail prompts across built-in, custom, extension, and MCP tools.
- Removed per-session run queues from JS and Python backends, allowing async cells on the same session id to interleave.
- Introduced `getEvalSessionId` on ToolSession so subagents spawned via `task` inherit the parent's executor id and share JS VM and Python kernel state.
- Switched JS runtime state from module-level fields to AsyncLocalStorage so concurrent runs route output and tool calls to their own context.
- Changed Python runner to an asyncio event loop with per-request tasks and ContextVar-based run id tracking for concurrent execution.
- Added mtime-based module cache eviction to preserve singleton state across re-imports of unchanged local files.
Combined task.enableLsp with the parent session enableLsp gate before spawning subagents, so --no-lsp remains authoritative even when subagent LSP is enabled in settings.
Fixes#1385
Added task.enableLsp (boolean, default false) and routed both regular and isolated subagent dispatch through it. Keeps subagents cheap by default while letting users opt in to LSP-aware delegation. Updated regression tests to cover the default-off, opt-in, plan-mode, and isolated paths.
Fixes#1385
Subagents now inherit the parent session's enableLsp value, so a top-level --no-lsp invocation propagates into spawned tasks instead of falling back to the executor's default of true.
Fixes#1385
Removed the hardcoded subagent LSP disable flag so executor defaults and user settings control LSP availability. Passed the effective plan-mode agent definition into both regular and isolated subagent dispatch so plan-mode tool restrictions apply consistently.
Fixes#1385
- Added `retry.maxDelayMs` to the settings schema and interfaces, with a default cap for provider backoff delays.
- Updated session auto-retry logic to fail fast when a requested wait exceeds the cap without fallback, emitting terminal auto-retry failure state.
- Propagated retry state and failure data into task progress and rendering so children show retry/wait details and reminder prompts stop after terminal errors.
Adds optional autoloadSkills field to agent frontmatter that automatically loads listed skills when a sub-agent is spawned. Uses the same buildSkillPromptMessage + sendCustomMessage mechanism as interactive skill loading, queued via sendCustomMessage({ triggerTurn: false }) before the first session.prompt(task). No extra agent turns, no new injection path. Skills stay in listing for sub-resource access. Compaction behavior matches manual loading. Unknown skill names silently skipped.
Lore-id: f85fdbdc
Constraint: autoload must use buildSkillPromptMessage + sendCustomMessage, never modify systemPrompt or use contextFiles
Constraint: triggerTurn must be false to avoid extra agent turns
Rejected: append to systemPrompt | agent cannot distinguish skill content from own instructions
Rejected: contextFiles injection | agent sees opaque file blob, cannot discover sub-resources
Rejected: promptCustomMessage per skill | N extra agent turns with model inference
Directive: autoload skill names are resolved against parent session skill list at spawn time in task/index.ts
Tested: TypeScript compiles clean with tsc --noEmit
Tested: parseAgentFields parses array and CSV string frontmatter
Tested: parseAgentFields returns undefined for absent and empty fields
Not-tested: bun test cannot run locally due to missing pi_natives native addon (requires Rust toolchain)
Confidence: high
Scope-risk: moderate
Reversibility: clean
Both empty-patch and whitespace-only-patch branches in the isolated-task
merge unconditionally appended "Applied patches: yes". The marker now
fires only when the combined patch is non-empty AND applyText succeeded.
No-op merges report "No changes to apply.". The renderer marker matcher
accepts the new string.
The "Started N background task jobs" announcement listed the per-task
label (e.g. 7-MyScout) without the underlying jobId, but JobTool.execute
looks up by jobId. Append (job: ${jobId}) to each task entry and accept
the task label as a fallback alias in JobTool.execute.
renderDescription filtered the listing only by disabledAgents. When a
subagent had parentSpawns empty the model still saw the full menu, tried
to spawn oracle, and got "Cannot spawn 'oracle'. Allowed: none". The
listing now intersects with the allowed-spawn set.
- Added defensive abort re-check after registering the abortSignal listener plus a checkAbort() immediately before await session.prompt(...), so a wall-clock timer that fires during pre-prompt setup is no longer lost between listener registration and the prompt call.
- Late yield events arriving after a wall-clock timeout no longer flip the result to success: a runtimeLimitExceeded flag derived from the internal abortReason forces wasAborted=true and exitCode=1 regardless of hasYield, while yield payloads remain captured in extractedToolData.
- Async task progress now copies contextTokens and contextWindow from the completed SingleResult onto AgentProgress, so backgrounded tasks still surface their context gauge to the UI.
- Added canonical `pi.zod` schema API exports and removed TypeBox package exports/imports.
- Migrated Tool schema typing from TypeBox to shared `TSchema`/Zod flow with legacy TypeBox compatibility.
- Updated AI provider adapters and MCP/agent builders to convert tool params through `toolWireSchema()`.
- Reworked schema validation from AJV to Zod-safe parsing with `fromTypeBox`, `toolWireSchema`, and meta schema checks.
- Task tool sessions now expose and forward parent OpenTelemetry config when creating subagent tasks.
- Subprocess execution now derives child telemetry from the parent config with the subagent identity and child session conversation handling.
- Subagent creation now records a handoff span using the resolved parent telemetry handle before running the child loop.
The async job completion path copied durationMs, tokens, and extractedToolData
from SingleResult back into the AgentProgress object, but missed cost.
Add progress.cost = singleResult?.usage?.cost.total ?? 0 to the same block.
Token counter (token_total status-line segment, subagent progress tree,
session-observer stats line) previously included cacheRead in its cumulative
sum. With Anthropic prompt caching, cacheRead per turn equals the full cached
context, so summing across N turns gives N*context_size -- a session with a 1M
context and 5 turns showed ~5M tokens despite no compaction occurring.
Fix: display shows input + output + cacheWrite per turn. cacheWrite is kept
because each byte is written once; cacheRead re-reads the same context every
turn. Dedicated cache_read/cache_write status-line segments still show cache
activity; billing cost is unaffected.
Also adds per-subagent cost display (dollar amount, statusLineCost color)
accumulated incrementally from message_end events. Hidden when cost is zero
(subscription/OAuth providers). Brings token and cost display in line with
what Claude Code shows per-agent.
- Updated the task tool output to list newly started background jobs by live task id with optional descriptions.
- Extended the async task prompt guidance to distinguish IRC-enabled versus standard coordination and cancellation behavior.
- Added unified isolation primitives (BackendKind, ProbeResult, IsoError) and resolve fallback selection logic.
- Added Diff, FileChange, and ChangeKind with default_diff choosing git mode or filesystem walk based on repository state.
- Added APFS/btrfs/zfs/reflink/overlayfs/projfs/rcopy/block-clone backends with platform-aware start, stop, and probe.
- Mapped backend operations to canonicalized paths, recursive clone helpers, rollback cleanup, and unavailable error mapping.
- Added unified native exports isoBackend/isoProbe/isoResolve/isoStart/isoStop/isoDiff and removed projfsOverlay APIs.
- Replaced task isolation resolver flow with ensureIsolation/cleanupIsolation and migrated mode handling to auto plus legacy-mode aliases.
- Added ownerId metadata to async jobs and to task/bash progress items from the session agent id.
- Extended async job registration and query methods with optional owner filters, and updated cancelAll to target matching owners.
- Updated session handoff and disposal so subagents inherit the parent manager, top-level sessions own it, and teardown cancels own jobs only.
- Added owner-aware async-job tests using hold/AbortSignal and scoped cancelAll assertions for running versus cancelled jobs.
- Passed the `irc.enabled` session setting into task prompt rendering so templates can branch for IRC mode.
- Updated orchestrator and subagent-task prompts with IRC-specific instruction paths for skip-all checks, coordination, and file-scoped assignments.
- Added parent-to-subagent artifact manager adoption so subagents reuse the parent `ArtifactManager` and write artifacts into a shared directory with shared IDs.
- Passed the shared artifact manager through tool/session context into subagent executor startup and exposed it via `SessionManager` and `ToolSession` for lookup.
- Updated kernel environment and artifact-resolution paths to prefer `PI_ARTIFACTS_DIR`, falling back to existing session-file-based behavior when absent.
- Added process-wide singleton instances for InternalUrlRouter, AsyncJobManager, and MCPManager.
- Changed internal URL protocols to resolve through registered sessions and scan all active roots/datasets for matches.
- Refactored agent, artifact, memory, rule, skill, jobs, and mcp handlers to use shared manager and rule/skill state.
- Removed per-session protocol/tool wiring and switched tests to initialize and reset global singleton state.
- Skipped writing compact conversation context files when IRC is enabled so subagents avoid using stale markdown snapshots.
- In runSubprocess, passed an undefined context file for IRC-enabled paths and kept context file prompting for non-IRC executions.
- Added explicit prompt markers and wrappers across system templates, including `[env]`, `[role]`, `[coop]`, `[closure]`, and `[now]`.
- Removed `renderTemplate` and `sectionSeparator` flows, deleted `task/template.ts`, and switched to per-task `renderSubagentUserPrompt` rendering.
- Updated system prompt assembly to `shortenPath`-normalize `cwd`, append rendered now metadata, and preserve trailing `[now]` blocks.
- Removed legacy template tests and added prompt-composition tests for ordered `[contract]`->`[project]`->`[now]` blocks and context-only system placement.
- Reworked shared prompt utilities by collapsing consecutive blank lines and removing obsolete `OPENING_HBS`/`LIST_ITEM` helper behavior.
- Added `listWorkspace` native binding and API types, exporting bounded workspace trees with AGENTS.md candidates.
- Reworked `buildWorkspaceTree` and `buildDirectoryTree` to call `listWorkspace` with 5s timeout defaults.
- Replaced startup AGENTS.md discovery with workspace-tree-only scanning and removed legacy AgentsMdSearch session plumbing.
- Updated `WorkspaceTree` and system prompt context to expose `agentsMdFiles` and aligned tests/changelog expectations.
Reporter screenshot showed a parent session on DeepSeek V4 Pro dispatching
a task subagent that resolved to `qwen3.6-plus-free` — an opencode-zen
model the user had no working credentials for. The dispatch hit a
provider that could not serve the model and surfaced a confusing API
rejection instead of using the parent's already-authenticated model.
Adds `resolveModelOverrideWithAuthFallback`, an auth-aware wrapper
around `resolveModelOverride` that checks the resolved subagent model's
credentials via `modelRegistry.getApiKey` + `isAuthenticated` and
falls back to the parent session's active model pattern when the
primary has no working auth. The parent's active model is plumbed
through `ExecutorOptions.parentActiveModelPattern` from `TaskTool`
into `runSubprocess`. If neither has working auth (or they resolve to
the same model), the primary resolution is preserved so the existing
error path still surfaces a meaningful failure downstream.
Fixes#985
Subagents previously re-ran buildAgentsMdSearch and buildWorkspaceTree on
every spawn, repeating the slowest part of system-prompt construction for
each task tool invocation. On large/pathological repos those scans
exceeded the 5s preparation deadline and tripped the per-subagent
'system prompt preparation timed out' warning.
Forward the parent's already-resolved AgentsMdSearch and WorkspaceTree
through createAgentSession (alongside the existing contextFiles, skills,
and promptTemplates inheritance):
- Add agentsMdSearch and workspaceTree to CreateAgentSessionOptions;
createAgentSession short-circuits the parallel scan promises when
these are provided.
- Resolve them with contextFiles before constructing ToolSession; expose
on ToolSession so the task tool can read the parent's values.
- Thread them through ExecutorOptions (task/executor.ts) into the
subagent's createAgentSession call, and pass them from the task tool
(task/index.ts) on both the worktree-isolated and non-isolated paths.
- Added optional `loadMode` and `summary` fields to `AgentTool` and related type declarations.
- Added `loadMode` and `summary` metadata to built-in tool classes for discoverable/essential behavior.
- Replaced `BUILTIN_TOOL_METADATA` with per-tool fields in discovery code paths.
- Updated `search_tool_bm25` and discovery indexing to use each tool's `summary` text.
- Updated discovery tests to validate tool `loadMode` and summary completeness.
- Added `HindsightSessionState` to `AgentSession` and bound hindsight lifecycle hooks to session state.
- Removed global hindsight state/queue handling and replaced it with per-session `HindsightRetainQueue` batching and scoped flushing.
- Reworked recall, reflect, and retain tools to use `session.getHindsightSessionState()` instead of sessionId-based lookup.
- Updated SDK/task/backend/controller flows to pass `session`/`parentHindsightSessionState`, scope `/memory` behavior, and document it in changelog.
- Added a localProtocolOptions field to session and executor option types for configurable local:// behavior.
- Passed localProtocolOptions through to LocalProtocolHandler creation and subagent session bootstrap.
- Propagated parent session local protocol settings from TaskTool so subprocess subagents share the same local:// artifacts and session context.
- Renamed the built-in `grep` content-search tool to `search` across settings, schemas, and SDK exports.
- Switched execution wiring so `Task`, `Plan`, cursor, and shell mapping now invoke `search` instead of `grep`.
- Updated prompts, plan-mode docs, and example tool lists to replace `grep`/`ls` references with `search` guidance.
- Aligned `Grep*`/`grep` event, renderer, and hook types to `Search*`/`search` across runtime and tests.
- Documented and fixed `search` result rendering budget behavior and added internal-URL/path-list transcript notes.
- Canonicalized file and CLI defaults from `read` to `open` across tool registration and prompts.
- Added `resolveToolAlias()` and applied alias-normalized tool selection so legacy `read` maps to `open`.
- Updated runtime, UI, and export layers to treat `open` as first-class while preserving `read` compatibility.
- Renamed read prompt docs to `open.md`/`open-chunk.md` and refreshed system guidance to recommend `open`.
- Updated tool-related tests and expectations from `read` to `open` (including test fixtures and aliases).
- Added task.simple to settings and schema with default, schema-free, and independent modes.
- Added mode-aware task schema and validation to enforce context/schema rules per simple mode.
- Updated prompts and template rendering to tailor headers and guidance for each simple mode.
- Added simple-mode capabilities and updated TaskTool execution for mode-aware context behavior.
- Added tests for independent rendering and mode-specific rejection of invalid context or schema inputs.
- Removed `SearchDb` APIs and `searchDb` fields, dropping db-backed state from native and agent sessions.
- Replaced crate export `fff` with `fd`, moving fuzzy-find bindings into `fd.rs`.
- Removed `SearchDb`/picker fast-path logic from `glob` and `grep`, simplifying scan flow and dropping db args.
- Removed `SearchDb`/`getSearchDb` wiring from extension, tool, and task context constructors across coding-agent.
- Added over-indentation validation warnings in chunk-edit normalization for suspicious `~` body line formatting.
- Removed `bytes`, `fff-grep`, `fff-search`, and `blake3` deps, adding `grep-searcher = "0.1"`.
- Added `vim` as an edit variant in benchmark CLI/config and rating script coverage.
- Expanded benchmark execution so `vim` is treated as a mutation tool for retries, stats, and edit intent checks.
- Adjusted `TaskTool` output schema precedence so explicit params override agent frontmatter.
- Fixed `TaskTool` success counting by excluding aborted tasks from success totals.
- Improved validation guidance in `SubmitResultTool`/`TodoWriteTool` for clearer recovery when payloads are missing or invalid.
- Added background command PID regression coverage in `executeBash` to confirm a real, terminateable PID is returned.
- Extracted prompt rendering and formatting utilities from coding-agent to centralized pi-utils package with new API surface (prompt.render, prompt.format, prompt.registerHelper).
- Migrated parseFrontmatter utility from coding-agent to pi-utils package; updated 8 files to import from @oh-my-pi/pi-utils.
- Removed 170-line prompt-format.ts module and consolidated 192 lines of Handlebars helper registrations into pi-utils prompt module.
- Updated 60+ files across coding-agent and typescript-edit-benchmark to use new prompt.render() and prompt.format() API from pi-utils.
- Simplified prompt-templates.ts by delegating core functionality to pi-utils while retaining custom helper registrations (jtdToTypeScript, jsonStringify, etc.).
Wire EventBus through task runtime so subagent lifecycle and progress
events propagate to the TUI. Add SessionObserverRegistry to track
active sessions, and a session observer overlay accessible via Ctrl+S
that shows a picker of running subagents and a read-only transcript
viewer that reads the subagent's session JSONL file to display
thinking, text, tool calls, and results.
- Add TASK_SUBAGENT_LIFECYCLE_CHANNEL for start/end events
- Add SubagentProgressPayload.sessionFile for session file tracking
- Pass eventBus through sdk.ts -> main.ts -> InteractiveMode
- SessionObserverRegistry with multi-listener onChange pattern
- Overlay picker preserves selection position across live refreshes
- Viewer renders full transcript from session JSONL (thinking, text,
tool calls with smart arg summaries, inline tool results)
- Extracted git operations from ControlledGit class into centralized utils/git module with 1276 lines of typed command wrappers.
- Replaced dependency injection of ControlledGit instances with direct cwd string parameters across commit agent tools and workflows.
- Migrated all git command execution from inline shell calls and custom helpers to structured git module API (diff, status, branch, worktree, patch, etc.).
- Removed ControlledGit class, operations.ts, and helper functions (findGitHeadPath, mergeStdoutStderr, joinPatch) now provided by git module.
- Exported git utilities from main package entry point for extension and plugin use.
- Consolidated fetch tool into read tool with URL reading capability and caching support.
- Removed standalone fetch tool from all agent prompts and CLI documentation.
- Extended read tool schema with timeout and raw parameters for URL fetch control.
- Added URL caching mechanism to prevent redundant network requests during read operations.
- Refactored fetch module from class-based tool to standalone executeReadUrl function.
- Updated read tool documentation to describe multi-purpose capabilities including web pages, GitHub, Stack Overflow, Wikipedia, Reddit, NPM, arXiv, blogs, and feeds.
- Added `wait_for_picker_scan()` function to search_db module with cancellation token support for polling picker scan completion.
- Integrated picker-based file search into glob matching logic with `collect_files_from_picker()` helper to reuse shared SearchDb picker results.
- Refactored fff and grep modules to use centralized `wait_for_picker_scan()` wrapper instead of direct FilePicker calls, improving cancellation handling.
- Propagated SearchDb instance through agent session initialization and input controller to enable unified file picker coordination across search operations.
can1357