opensec-pi-subagents 0.20.3 → 0.20.5

package/CHANGELOG.md

@@ -7,6 +7,97 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.20.5] - 2026-10-16
+
+Orchestration results the main agent can act on: failed runs are error results with what they produced, unfinished
+agents point at the completion notification instead of inviting polling, and a released agent's result stays readable.
+
+> **Possibly breaking:**
+> - **Failed, aborted and stopped runs return `isError: true`**, as do unknown-agent lookups and refused steers or
+>   resumes. Pi shows them as tool errors and the main agent reads them as failures.
+> - **A model name whose closest matches span several providers is refused** instead of resolving to one of them,
+>   unless the parent's provider is among them. Use the exact `provider/id` to choose another provider.
+
+### Added
+
+- **`get_subagent_result` reads a released agent's result from its saved session.** After the 10-minute cleanup it
+  used to answer `Agent not found`; it now answers with the status and model from the kept resume state and the final
+  result read from the session file, marked as such. Nothing is held in memory longer: the file is read on demand. The
+  resume state now records the run's status.
+- **Each run records the model and thinking level it ran with**, a resumed run its own. The foreground result's
+  stats, `get_subagent_result` (`Model:`) and the completion notification (`<model>`, `<thinking>`) show them.
+- **Malformed tool arguments from weak models are repaired before validation** (`prepareArguments`) on the top-level
+  and nested `Agent`, `get_subagent_result` and `steer_subagent` tools: the arguments object sent as a JSON string or
+  stuffed into one string field, arrays where a string is expected, `null` optional fields, booleans and numbers sent
+  as strings, and a missing `subagent_type` on the top-level `Agent` tool, which goes to the default type under the
+  `fallbackSubagent` policy (`Note: No agent type given — using …`). A missing prompt is never invented.
+
+### Changed
+
+- **Failed runs are error results.** The foreground `Agent` result, resumed runs and `get_subagent_result`, top-level
+  and nested, set `isError` for `error`, `aborted` and `stopped`. The result renderer keeps the agent's status line for
+  these. A stopped run's headline reads `Agent was stopped by the user after …`, an aborted one `Agent was aborted
+  after …`, instead of `Agent completed in …`.
+- **Failure results keep the partial output and name the session file.** The run's last assistant text is kept also
+  when the run threw or a stopped run was abandoned, bounded to the run's own turns, and the result names the agent's
+  saved session file, which holds every turn (`<session-file>` in the notification). No separate files are written.
+- **No more polling prompts.** `get_subagent_result` on a running, queued or still-stopping agent says a completion
+  notification arrives on its own, to continue with other work, and not to call it again for that agent until then.
+  Its description and the `wait` parameter say the same; `wait` stays `false` by default. Background spawn and resume
+  results and the refused resume of a running agent no longer suggest `get_subagent_result` to wait.
+- **Notifications preview the first 10 lines of a result**, still capped at 500 characters (300 per agent in a group)
+  so one long line cannot flood the context, and carry the full result's size in `<result_bytes>`. The truncation note
+  says how many lines and bytes were left out.
+- **Model names no longer resolve across providers by guessing.** An exact `provider/id` (case-insensitive) resolves
+  as before. Among a short name's closest matches (`haiku`, `sonnet`), the parent's provider wins when it serves one;
+  matches from a single provider resolve to the tightest, as before (ties by id). Only matches that still span several
+  providers are an error naming up to five closest candidates. An exact id two providers serve follows the same rule.
+  A per-call model that does not resolve stays a tool error on every path.
+- **An agent file's model that does not resolve is reported.** The run still falls back to the parent's model, but its
+  results say which model was asked for, why it was not used and which model ran instead (the `Agent` and nested
+  `Agent` results, `get_subagent_result`, `<model-note>` in the notification). Spawns that resolve the agent file's
+  model by exact `provider/id` only (`@mentions`, RPC, scheduled jobs) and scheduled jobs whose model no longer
+  resolves note it the same way; workflows show it as a warning once per run.
+- **More released agents stay addressable:** 2000 top-level and 4000 nested or workflow entries (was 100 and 200).
+  An entry holds ids, paths and settings, about 450 bytes (700 with a worktree), never the conversation. Restoring
+  them at session start trims once instead of rescanning per entry.
+
+## [0.20.4] - 2026-10-16
+
+The conversation viewer draws a subagent's transcript with Pi's own chat components, so it looks exactly like Pi's
+main session.
+
+> **Possibly breaking:** the `viewerMarkdown` setting (`/agents → Settings → Viewer markdown`) and the viewer's `m`
+> key are removed: Pi's components decide how each message renders. A `viewerMarkdown` key left in `subagents.json`
+> is ignored.
+
+### Changed
+
+- **The conversation viewer (regular overlay and fullscreen observer) renders like Pi's main session.** It builds
+  `UserMessageComponent`, `AssistantMessageComponent` (thinking and text), `ToolExecutionComponent` (each tool's own
+  call and result renderers from the subagent's session: edit diffs, read previews, bash command, output, exit status
+  and duration, running/done/error state, arguments), `BashExecutionComponent`, `CompactionSummaryMessageComponent`,
+  `BranchSummaryMessageComponent`, `SkillInvocationMessageComponent` and `CustomMessageComponent` from the session's
+  messages as Pi's interactive mode does, with the same spacing and collapsed defaults. Streaming messages, tool
+  progress and results update live. It replaces the viewer's own `[User]` / `[Assistant]` / `[Tool: name]` /
+  `[Result]` rendering and its 16k-character result cap (Pi's renderers truncate collapsed output themselves).
+- **`ctrl+o` expands and collapses tool output in the viewer**, as in Pi; it follows a rebinding of Pi's
+  `app.tools.expand`, and the footer shows it when there is room. Pi's `hideThinkingBlock`, `outputPad` and
+  `markdown.codeBlockIndent` settings apply.
+- Each component is built once and its rendered lines are cached per width and expand state; a frame re-renders only
+  the streamed message and tools that changed, scrolling never re-renders, and only visible rows are assembled.
+  Without a JIT (Bun with `BUN_JSC_useJIT=0`, standing in for Pi-Bolt), at 500 messages: a frame 6 µs, a scroll step
+  15 µs, a streaming delta 0.70 ms (was 0.92 ms), a tool call finishing 0.26 ms (was 0.31 ms), the first frame 121 ms
+  (was 160 ms). A new user message costs 0.27 ms (was 0.17 ms): Pi's user message is Markdown on a background.
+- Where the viewer differs from Pi's main session: images in tool results show as Pi's text placeholder, since an
+  inline image cannot be clipped to a scrolled overlay; Mermaid blocks are not drawn as diagrams (Pi's Mermaid
+  transformer is not exported); OSC 133 prompt marks are not emitted inside the overlay; clicking a tool or thinking
+  block does not toggle it (use `ctrl+o`).
+
+### Removed
+
+- The `viewerMarkdown` setting and the viewer's `m` key (see above).
+
 ## [0.20.3] - 2026-10-16
 
 Background agents survive a session switch, subagents run on the parent's model by default, and nested and workflow

package/README.md

@@ -117,6 +117,8 @@ locations are those of `@tintinweb/pi-subagents`.
 
 Based on `@tintinweb/pi-subagents` 0.19:
 
+- **0.20.5:** failed, aborted and stopped agents come back as error results with their last output and session file; weak models' malformed tool arguments are repaired; a running agent's reply asks the model to wait for its notification instead of polling; notifications preview 10 lines and give the result's size; each run's model is recorded and shown; an ambiguous model name is no longer settled by a guess across providers.
+- **0.20.4:** the conversation viewer draws a subagent's transcript with Pi's own chat components, as Pi's main session does; `ctrl+o` expands and collapses tool output. The `viewerMarkdown` setting and the `m` key are gone.
 - **0.20.3:**
   - From upstream 0.20–0.22: completions reach a busy parent before its next model call; an agent file's `isolated:`/`isolation:` hold on every spawn path; tool scoping covers tools called from other tools; `ctrl+b`; the fullscreen viewer, built on the incremental viewer (a frame in about 10 µs at 500 messages without a JIT).
   - Nested and workflow agents are resumable after a restart; background agents keep running across `/new`, `/resume`, `/fork` and `/reload`, and their results reach the session that started them.