Replace flaky time-based heuristic (sleep + 500ms race) with a proper
ready-signal protocol. The RPC server now emits {"type":"ready"} on
stdout when initialized, and the client waits for that signal instead
of guessing based on timing.
Fixes CI failure where the process took longer than 600ms to reach
provider validation (due to auth discovery + model registry refresh),
causing start() to resolve even when the process was about to exit.
- Added time-budget-based retry strategy for Google Gemini 429 rate limit errors with 5-minute window.
- Set default maxRetries to 5 for Anthropic, Azure OpenAI, and OpenAI clients.
- Implemented separate 429 rate limit handling with cumulative time budget tracking instead of bounded retry attempts.
Split Codex error handling: usage_limit_reached/usage_not_included show
"usage limit" message, while rate_limit_exceeded/bare 429 show "rate
limit exceeded" — so transient throttling no longer looks like a
subscription cap.
Changed 429 retry strategy: when the server provides a retry delay
(retry-after header or body hint), use a 5-minute time budget instead
of the fixed 5-attempt cap. Transient rate limits now wait and retry
rather than giving up after a few seconds.
- Improved OAuth token refresh error messages to include provider-specific error details from API responses.
- Separated rate limit and usage limit error handling in OpenAI response handler with distinct error messages and retry timing.
- Enhanced error message propagation in OAuth refresh flow to preserve original error reasons for better debugging.
- Fixed regex pattern in auth storage to use word boundaries for accurate HTTP status code matching.
- Added `includeDisabled` parameter to `listAuthCredentials()` to optionally retrieve disabled credentials.
- Added `disableAuthCredential()` method for soft-deleting auth credentials while preserving database records.
- Changed auth credential removal to use soft-delete (disable) instead of hard-delete when OAuth refresh fails, keeping credentials in database for audit purposes.
- Added `disabled` column to auth_credentials table schema with automatic migration from v3 to v4.
- Added prepared statements for querying active (non-disabled) credentials with optional provider filtering.
- Add 'prepack' script so bun publish regenerates the index before packing
- Add root-level 'prepare' script so bun install regenerates after cloning
- Gitignore and untrack docs-index.generated.ts
- Remove 'docs' from package.json 'files'
fixes#116
Static imports from '../../../../docs/' resolve correctly in the monorepo
but break when the package is installed from npm -- the docs/ directory
at the repo root is not included in the published tarball.
Switch the generator to read each file at build time and emit the content
as inlined string literals, eliminating all runtime file resolution.
- Added docs:// internal URL protocol handler for accessing embedded markdown documentation files with path traversal validation.
- Added support for executable tool files (.ts, .js, .sh, .bash, .py) in custom tools discovery from .omp/tools/ and .claude/tools/ directories.
- Added generate-docs-index build script to automatically scan and embed markdown documentation files at build time.
- Updated system prompt to document docs:// protocol usage and identify agent as operating within Oh My Pi harness.
- Changed hashline format separator from pipe (|) to colon (:) for improved readability across all tools and output formats.
- Refactored hashline edit API with operation-based structure: renamed delete->rm, rename->mv, set->target/new_content, and added explicit op field for operation types.
- Updated hashline hash encoding from 4-character base36 to 2-character hexadecimal for more compact representation.
- Replaced anchor terminology with tags throughout hashline documentation and API for clearer semantics.
onAutocompleteUpdate was calling requestRender(true), which resets all
differential rendering state on every keystroke while the slash menu is
open (debounced at 100ms). This caused the TUI to perform a full screen
clear and repaint on each filter update, producing visible flicker when
the editor is pushed to the bottom of the terminal.
onAutocompleteCancel retains requestRender(true) since it fires only
once on dismissal and the forced redraw is needed to erase the stale
menu rows that the differential renderer would otherwise leave on screen.
onAutocompleteUpdate is changed to requestRender() (no force), letting
the differential renderer repaint only the changed lines within the
autocomplete area.
- Added `gemini-3.1-pro` model to opencode provider with text and image input support.
- Added `trinity-large-preview-free` model to opencode provider.
- Added `google/gemini-3.1-pro-preview` model to nanogpt provider.
- Added `google/gemini-3.1-pro-preview` model to openrouter provider with text and image input support.
- Added `gemini-3.1-pro` model to cursor provider.
- Changed `big-pickle` and `minimax-m2.5-free` models from `openai-completions` to `anthropic-messages` API and updated baseUrl endpoints.
- Added streamed tool intent display in working message to show real-time intent tracking during agent execution.
- Changed intent tracing field name from `$intent` to `_intent` across tool schemas and agent core for consistency.
- Added support for file deletion and renaming operations in hashline edit mode.
- Renamed hashline edit operation fields: `set` to `target`/`new_content`, `set_range` to `first`/`last`/`new_content`, `insert` to `inserted_lines`.
- Added optional `intent` field to `ToolCall` interface for capturing harness-level intent metadata.
- Added `intentTracing` configuration option to enable intent goal extraction from tool calls with automatic `$intent` field injection and argument stripping.
- Implemented intent injection and extraction logic in agent-loop to populate tool call intent metadata when intentTracing is enabled.
- Added `tools.intentTracing` setting to coding-agent configuration schema with environment variable override support.
- Added support for resolving internal artifact:// URLs in grep tool to search backing files.
- Fixed grep tool to properly handle internal URL resolution with validation for missing backing files.
- Added comprehensive test suite covering artifact URL resolution, regex patterns, and error handling.
- Optimized CI matrix to conditionally include platform variants based on git tag presence.
- Redesigned hashline edit API with new operation names (set, set_range, insert) and structured body parameter accepting string arrays for multiline edits.
- Changed hashline reference format from LINE:HASH to LINE#ID throughout tools and documentation for improved clarity.
- Enhanced insert operation to support optional before/after anchors enabling flexible insertion positioning and boundary echo stripping.
- Made hashline autocorrect heuristics conditional on PI_HL_AUTOCORRECT environment variable for controlled behavior.
- Added benchmark reports for claude-haiku-4-5 and GPT-5.2-Codex models demonstrating hashline edit variant performance.
- Added flushOutput() public method to bash-interactive component for synchronizing terminal output completion.
- Enhanced SSH host discovery to load from managed omp config paths (.omp/ssh.json and ~/.omp/agent/ssh.json) with deduplication.
- Improved terminal output handling in interactive bash sessions to ensure queued writes complete before processing pending chunks.
- Refactored write queue consumption in bash-interactive to use index-based tracking instead of array shift operations.
- Added SSH host management feature with `ssh` CLI command and `/ssh` slash command for add, list, and remove operations.
- Added SSH configuration support at project (.omp/ssh.json) and user (~/.omp/agent/ssh.json) scopes with host discovery from project files.
- Added SSH host configuration flags: --host, --user, --port, --key, --desc, --compat, and --scope for flexible host setup.
- Removed automatic line relocation on stale hash references; now fails with error instead.
- Added provider failure detection and exponential backoff retry logic to handle authentication and authorization errors in benchmark tasks.
- Implemented HashlineMismatchError behavior in coding-agent to fail on stale hash references instead of silently relocating edits.
- Simplified hashline validation by removing automatic line relocation logic and hash tracking infrastructure.
- Added benchmark report for claude-sonnet-4-6 model showing 85% task success rate with detailed failure analysis and performance metrics.
- Fixed terminal output handling to properly queue and serialize writes, preventing dropped or corrupted output in interactive bash sessions.
- Implemented write queue with callback-based draining to ensure sequential terminal writes.
- Removed try-catch fallback in output handling, relying on queued serialization for correctness.
- Added 'unable to connect' to transient error patterns in retry logic to properly handle connection failures.
- Updated retry error detection in coding-agent to match ai package pattern for consistency.
