- 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.
- Added helpers to stringify and truncate display() JSON values before including them in text responses.
- Updated eval execution to append formatted display output text alongside stdout and emit images as dedicated content blocks.
- Adjusted no-output messaging for image-only runs and removed detail-level image payload duplication when content already includes image blocks.
- Updated bash, job, and task prompts to state that background results are delivered automatically when complete.
- Removed guidance encouraging repeated `jobs://` polling and clarified `job` with `poll` should be used only when a task is blocking.
- Changed the bash tool confirmation message to recommend doing other work while waiting for background jobs and polling only if needed.
- Removed the shared date, workspace, and critical-response block from now-prompt.md.
- Added the same instructional block to project-prompt.md so the project prompt now carries those system constraints.
- Updated `irc` tool availability logic to require both `irc` settings and concurrency support.
- Imported `MAIN_AGENT_ID` and suppressed IRC access for the main agent when async mode is disabled, preventing sync-mode peer messaging.
- Replaced `===== ... =====` eval cell headers with `*** Begin ` / `*** End ` markers; legacy format remains renderable in HTML exports.
- Replaced hashline patch grammar with `*** Begin Patch` / `*** End Patch` envelope; old inputs without the envelope are still accepted.
- Extracted `sniffEvalLanguage` into a shared `sniff.ts` module reused by the parser and tool.
- Added `docs/ERRATA-GPT5-HARMONY.md` and `scripts/session-stats/harmony_backtest.py` documenting and backtesting the GPT-5 Harmony-header leak defect.
- Deleted `bash-normalize.ts` and its tests; output truncation is handled by the streaming tail buffer and artifact spillover.
- Removed `head`/`tail` schema fields from `bashSchema` and `BashToolInput`.
- Updated system prompt and bash tool prompt to forbid `| head`/`| tail` pipes and other anti-patterns, directing the agent to use dedicated tools instead.
- Added immutable metadata to protocol handlers and had the router copy each handler's setting onto resolved internal resources.
- Updated read and search tools to honor `resource.immutable`, and made search suppress hashline anchors only for immutable source paths while preserving them for mutable files.
- Expanded internal URL tests to cover mutable local resources and mixed immutable/mutable search input hashline behavior.
- Extended splitPathAndSel to detect compound selectors that combine a line range with `raw` in either order.
- Updated read selector parsing to support `:lines:raw` and `:raw:lines` selectors and propagate raw-mode handling through internal URL, archive, and notebook paths.
- Added tests covering compound selector syntax and confirming it returns verbatim content without anchors or line-number prefixes.
- Extended file-display resolution to accept an immutable flag and suppress hashline anchors when immutable sources are requested.
- Plumbed immutable propagation through ReadTool and SearchTool so internal URL reads and searches pass that context.
- Added a regression test confirming artifact:// search output no longer includes hashline anchors.
- 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.
- Added `./hashline` package exports and redirected callers to the new hashline entrypoint.
- Moved hashline logic out of `edit/` to `src/hashline` and removed `edit/modes/hashline`/`edit/line-hash` paths.
- Added hashline parsers, anchors, types, and diff helpers with stricter input and mismatch validation.
- Implemented preflight and cache-recovery execution flows to reapply edits and handle stale anchor mismatches.
- Documented the hashline API relocation as breaking changes in `CHANGELOG.md`.
- Implemented hashline stale-anchor recovery using cached reads and a 3-way merge fallback.
- Updated hashline execution and preflight checks to retry mismatched anchors through cache recovery.
- Added file-read cache support with per-session LRU snapshots and contiguous/sparse record APIs.
- Integrated read and search tools with the shared cache to record candidate lines for recovery.
- Added tests for stale-anchor recovery flows and FileReadCache session, null, overlap, and eviction behavior.
- Expanded read tool range calculations to include optional leading and trailing context lines around user-requested offsets and limits.
- Added shared context range expansion logic so line-slice and streaming reads return anchor-safe windows while preserving existing truncation behavior.
- Updated read-tool tests to assert boundary-offset, limit, and combined offset-limit reads now include the expected ±3 lines of context.
- Added an intent function to EvalTool that generated a label from provided input.
- Parsed input with parseEvalInput and joined each cell title or language fallback into a newline string.
- Returned "evaluating" when input was missing or could not be parsed.
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.
- Updated the read tool docs to define URL selectors as `:50`, `:50-100`, and `:50+150` and to document the `https://host/:port` form for URLs with explicit ports.
- Changed URL selector parsing to use the file-style numeric range format, treated `raw` as a separate token, and updated the invalid zero-line message to direct users to `:1`.
- Refined embedded URL selector extraction to validate the base URL first and then match selectors against the new numeric range regex.
- Removed the read CLI argument and tool schema field so read requests no longer accept custom timeouts.
- Updated URL read handling to stop forwarding timeout values and execute URL reads without a timeout parameter.
- Standardized URL read fetching to a fixed 30-second timeout and dropped timeout metadata from URL call rendering.
- Added a readonly intent property to the write tool implementation.
- The intent now returned a contextual message using `args.path` when available, otherwise a generic "writing" label.
- Adjusted streaming preview formatting to render the tail section using a computed start index and hidden-line count.
- Passed language into streaming formatting to syntax-highlight visible lines and include aligned line-number gutters.
- Updated the hidden-lines notice to handle singular vs plural earlier-line counts correctly.
- Updated the write tool preview renderer to syntax-highlight visible lines and annotate them with line-number gutters.
- Changed the file metadata output to an inline line-count suffix on the header and removed the separate metadata row.
- Adjusted preview truncation to show the first lines first while preserving hidden-line hints for collapsed output.
- Added `.ipynb` detection and editable-cell conversion utilities, including merge/serialize helpers in `edit/notebook`.
- Rerouted hashline, patch, and replace edit flows, plus read/write paths, through notebook-aware helpers before persistence.
- Removed the dedicated `notebook` tool, its schema flags, renderer, and built-in registration/settings checks.
- Updated notebook read behavior, docs, and tests so `.ipynb` reads return editable `# %%` cells and edits reserialize to JSON.
- 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.
Restore BUILTIN_TOOLS to Record<string, ToolFactory> so external SDK callers can
still invoke BUILTIN_TOOLS.read(session) directly, and move per-tool discovery
metadata (loadMode, summary) into a dedicated BUILTIN_TOOL_METADATA map. All
internal callers (computeEssentialBuiltinNames, getBuiltinDiscoverableEntries,
createTools, sdk.ts initial-tool filter, agent-session built-in collection) now
read metadata through the new map.
Restore the legacy MCP discovery API on AgentSession: getDiscoverableMCPTools()
returns DiscoverableMCPTool[] with description, and getDiscoverableMCPSearchIndex()
returns the legacy DiscoverableMCPSearchIndex whose documents expose
tool.description while remaining usable by searchDiscoverableTools (summary is
populated from description so the BM25 corpus still scores correctly). Generic
discovery via getDiscoverableTools / getDiscoverableToolSearchIndex is unchanged.
Centralize discovery cache invalidation in #invalidateDiscoveryCaches and call it
from #applyActiveToolsByName, refreshMCPTools, and refreshRpcHostTools so the
generic search index can no longer return tools that have already been activated
or registry entries that have been replaced.
Restrict #collectDiscoverableBuiltinTools to entries whose
BUILTIN_TOOL_METADATA[name].loadMode === "discoverable", which keeps hidden
tools (resolve, yield, exit_plan_mode, report_finding, report_tool_issue) and
unknown extension/custom registry entries out of the discovery corpus.
Tests: add coverage for callable BUILTIN_TOOLS factories, legacy MCP description
shape on getDiscoverableMCPTools / getDiscoverableMCPSearchIndex, stale-index
invalidation on setActiveToolsByName, and hidden-tool exclusion from
getDiscoverableTools({ source: "builtin" }).
The internal class is now SearchToolsTool but the wire name stays
"search_tool_bm25" so persisted MCP selections in user session files
continue to resolve correctly.
The BM25 corpus is now drawn from the generic DiscoverableTool list
(built-ins, MCP, extension, custom) rather than only MCP tools.
createIf fires when either tools.discoveryMode !== "off" or the
legacy mcp.discoveryMode is true.
AgentSession gains generic discovery methods (isToolDiscoveryEnabled,
getDiscoverableTools, getDiscoverableToolSearchIndex,
getSelectedDiscoveredToolNames, activateDiscoveredTools). The existing
MCP-named methods are kept as thin shims that filter by
source === "mcp" so existing callers continue to work.
Updated the tool's prompt copy to advertise discovery across all
sources rather than MCP only. Tests extended for the new shape.
Added two settings:
- tools.discoveryMode ("off" | "mcp-only" | "all", default "off")
- tools.essentialOverride (string[], default empty)
Converted BUILTIN_TOOLS from Record<string, ToolFactory> to
Record<string, BuiltinEntry> with a per-tool loadMode ("essential"
or "discoverable") and an optional summary used as the BM25 corpus
entry when discovery hides the tool.
Marked read, bash, edit as essential. Marked the remaining 24 built-in
tools as discoverable with hand-written summaries. search_tool_bm25
stays essential (always loaded when discovery is on; gated separately
by isToolAllowed).
Added DEFAULT_ESSENTIAL_TOOL_NAMES, computeEssentialBuiltinNames
(reads tools.essentialOverride with a default fallback), and
getBuiltinDiscoverableEntries (used by the search tool to build the
BM25 corpus).
mcp.discoveryMode is preserved as a back-compat alias for "mcp-only".
- Added a new `read.summarize.prose` setting to control markdown and plain-text read summaries.
- Updated the read tool to skip summarization for prose files unless that setting is enabled.
- Added coverage for default and enabled markdown read-summary behavior.
- Exported DEFAULT_MATCH_LIMIT from the search tool module.
- Updated the search tool tests to import DEFAULT_MATCH_LIMIT from the module.
- Updated default-limit assertions and messages to use DEFAULT_MATCH_LIMIT instead of hardcoded values.
- Added `pr_create` to the GitHub tool schema and execution dispatch.
- Implemented `executePrCreate()` with title/fill validation, arg assembly, and temp-file body handling.
- Added optional PR result enrichment by parsing created PR URLs and formatting `gh pr view` summary fields.
- Updated `github.md` docs, changelog notes, and tests for `pr_create` mappings and validation.
- Converted systemPrompt APIs and state types to ordered `string[]` across agent, AI, and coding-agent surfaces.
- Added `normalizeSystemPrompts` and applied it to context normalization before building provider request payloads.
- Updated AI providers to emit separate normalized prompt blocks/messages instead of a single merged system prompt.
- Removed dedicated `projectPrompt` state and remapped that context into system-context buckets in session, dump, and token accounting.
- Aligned tests and changelogs to pass and assert `systemPrompt` as arrays with ordered prompt semantics.
- Added DirectoryTree, DirectoryTreeOptions, and buildDirectoryTree exports for configurable tree rendering.
- Changed read tool directory output to use buildDirectoryTree with depth and exclusion limits.
- Added read.summarize settings and summary-mode parseable read output behavior when no selector is used.
- Added tests for truncated root/child listings and hidden or excluded entry filtering.
- `is_elidable_kind` now treats more AST kinds as elidable for TypeScript, Rust, and Java.
- Added summary tests for interface and class bodies in TypeScript, Rust, and Java.
- Merged brace-boundary read summary spans into a single `..` line-range representation.
- Grouped adjacent groupable nodes and emitted elision spans only after grouped runs met thresholds.
- Normalized end-line rows at column 0 and extended import-like groupable-kind rules across languages.
- Updated read-summary tests and added changelog notes for merged brace-pair `{ .. }` output behavior.
- Updated read schema, path utilities, and dispatch to parse selectors from :raw/:L suffixes on path, removing standalone sel usage.
- Changed truncation/error notices to continue with :<nextOffset> and :1 guidance for read and sqlite pagination.
- Added summarizeCode support with tree-sitter summaries, read.summarize settings, and N-API Summary types/exports.
- Added tests and docs updates for path-embedded selectors, summary behavior, and explicit raw/offset SQL/read cases.
- Updated OpenAI completion request construction to include OpenRouter attribution headers and always-on cache directives (1h TTL).
- Updated image-generation requests to replace legacy X-Title with OpenRouter-specific HTTP-Referer, title, and category headers.
- Removed old OpenRouter X-Title defaults from provider options and removed the matching image-gen header assertion test.
- 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.
- Loaded recall, reflect, and retain tool descriptions from new markdown prompt files.
- Updated the retain tool to accept batched `items`, enqueue each memory entry, and return the queued-memory count.
- Reworked retain tests to pass the array payload, assert pluralized queue messages, and confirm single-batch flush behavior.
- Added per-session retain queues with size/time auto-flush, recursive drain, and lifecycle flushes on end/clear/enqueue.
- Changed `hindsight-retain` to validate session state, enqueue writes, and return `Memory queued.` immediately.
- Added `notice` event support and handlers that route error/warning/status messages with source-aware formatting.
- Updated client and tests with shared request mapping, `RequestOptions`, `buildMemoryItem`, and expanded batch/list/doc APIs.
- Updated tool selection logic to add memory-specific tools when the session backend is set to "hindsight".
- Ensured the new tools "recall", "retain", and "reflect" are appended only if not already requested.
- Added inline hashline parse and apply support for `<` prepend and `+` append operations with prefix+suffix edits.
- Added fail-fast behavior to reject inline modify ops combined with delete or replace on same line.
- Renamed HASHLINE_* and mode symbols to HL_* in prompt tooling, read/search checks, and prompt templates.
- Standardized separators to `PI_HL_SEP`/`HL_EDIT_SEP` and fixed `HL_BODY_SEP='|'`, updating parser formatting behavior.
- Updated benchmark subtype constants and python cleanup test setup to use HL_* values and AgentRegistry mock failure injection.
- Added `hindsight.scoping` and `HINDSIGHT_SCOPING` with global, per-project, and per-project-tagged modes.
- Migrated legacy `raw.hindsight` handling to scoping-based behavior and removed deprecated `dynamicBankId`/`agentName` keys.
- Replaced `deriveBankId` with `computeBankScope`, returning `bankId` plus optional retain/recall tags by mode.
- Forwarded scoped tags through memory ops so recall and retain/reflect calls use `tags`, `tagsMatch`, and `recallTags`.
- Validated `scoping` values from env/settings, warning and defaulting to `per-project-tagged` when invalid.
- Standardized Hindsight context output to emit `<memories>` blocks and updated retention stripping to remove both `<memories>` and legacy `<hindsight_memories>`/`<relevant_memories>` tags.
- Renamed built-in Hindsight tool entry points and labels from `hindsight_retain`/`hindsight_recall`/`hindsight_reflect` to `retain`, `recall`, and `reflect`, including tool availability checks.
- Added a new optional `beforeAgentStartPrompt` hook to `MemoryBackend` and implemented it in the Hindsight backend to recall long-term context for the first turn.
- Updated `AgentSession` startup flow to inject the recalled context into the turn-specific system prompt before the first response is generated.
- Preserved `<hindsight_memories>` tags in Hindsight developer instructions and added tests for first-turn injection and state caching.
- Added `memory.backend` and `hindsight.*` settings schema with migration from `memories.enabled` legacy mode.
- Added Hindsight memory backend runtime modules for resolved config, client creation, bank ID derivation, and state lifecycle.
- Added off/local/hindsight backends and resolver wiring across SDK, commands, and compaction context.
- Added `hindsight_recall`, `hindsight_reflect`, and `hindsight_retain` tools with schema validation and backend gating.
- Added Memory tab metadata and symbols to expose backend selection in the settings UI.
- Added package export barrels and tests for bank ID, content formatting, and hindsight config env precedence.
- Updated search, find, ast-edit, and ast-grep to skip missing paths and run on existing ones.
- Added multi-path error handling to throw ToolError only when all provided paths are missing.
- Updated search and find outputs to surface skipped `missingPaths` in text and renderer warnings.
- Updated CHANGELOG with multi-path tool behavior updates and `search_repos` global-search notes.
- Added `PartitionedPaths` and `partitionExistingPaths` to classify existing versus missing path inputs.
- Added test helpers and fixtures for multi-path missing-path cases in search/find behavior tests.
can1357