@oh-my-pi/pi-coding-agent 15.5.13 → 15.6.0

package/src/modes/theme/theme.ts

@@ -144,6 +144,7 @@ export type SymbolKey =
 	| "md.quoteBorder"
 	| "md.hrChar"
 	| "md.bullet"
+	| "md.colorSwatch"
 	// Language/file type icons
 	| "lang.default"
 	| "lang.typescript"
@@ -308,6 +309,7 @@ const UNICODE_SYMBOLS: SymbolMap = {
 	"md.quoteBorder": "▏",
 	"md.hrChar": "─",
 	"md.bullet": "•",
+	"md.colorSwatch": "■",
 	// Language/file icons (emoji-centric, no Nerd Font required)
 	"lang.default": "⌘",
 	"lang.typescript": "🟦",
@@ -568,6 +570,8 @@ const NERD_SYMBOLS: SymbolMap = {
 	"md.hrChar": "─",
 	// pick:  | alt:  •
 	"md.bullet": "\uf111",
+	// pick: ■ | alt:  (U+F096)
+	"md.colorSwatch": "■",
 	// Language icons (nerd font devicons)
 	"lang.default": "",
 	"lang.typescript": "\u{E628}",
@@ -730,6 +734,7 @@ const ASCII_SYMBOLS: SymbolMap = {
 	"md.quoteBorder": "|",
 	"md.hrChar": "-",
 	"md.bullet": "*",
+	"md.colorSwatch": "[]",
 	// Language icons (ASCII uses abbreviations)
 	"lang.default": "code",
 	"lang.typescript": "ts",
@@ -1519,6 +1524,7 @@ export class Theme {
 			quoteBorder: this.#symbols["md.quoteBorder"],
 			hrChar: this.#symbols["md.hrChar"],
 			bullet: this.#symbols["md.bullet"],
+			colorSwatch: this.#symbols["md.colorSwatch"],
 		};
 	}
 
@@ -2340,6 +2346,7 @@ export function getSymbolTheme(): SymbolTheme {
 		table: theme.boxSharp,
 		quoteBorder: theme.md.quoteBorder,
 		hrChar: theme.md.hrChar,
+		colorSwatch: theme.md.colorSwatch,
 		spinnerFrames: theme.getSpinnerFrames("activity"),
 	};
 }

package/src/modes/ultrathink.ts

@@ -1,5 +1,5 @@
 import ultrathinkNotice from "../prompts/system/ultrathink-notice.md" with { type: "text" };
-import { theme } from "./theme/theme";
+import { createGradientHighlighter } from "./gradient-highlight";
 
 /**
  * "ultrathink" keyword support, mirroring Claude Code's affordance.
@@ -11,12 +11,8 @@ import { theme } from "./theme/theme";
  * "ultrathinking"/"ultrathinks" never trigger either behavior.
  */
 
-// Cheap, stateless presence probe used to skip the boundary regex on most lines.
-const ULTRATHINK_PROBE = /ultrathink/i;
 // Detection: standalone keyword, any case. Non-global so `.test` stays stateless.
 const ULTRATHINK_WORD = /\bultrathink\b/i;
-// Highlight: global so `.replace` walks every occurrence.
-const ULTRATHINK_HIGHLIGHT = /\bultrathink\b/gi;
 
 /** Hidden system notice appended after a user message that mentions "ultrathink". */
 export const ULTRATHINK_NOTICE: string = ultrathinkNotice.trim();
@@ -26,54 +22,14 @@ export function containsUltrathink(text: string): boolean {
 	return ULTRATHINK_WORD.test(text);
 }
 
-const FG_RESET = "\x1b[39m";
-// Hue stops swept across the visible spectrum. More stops than the keyword has
-// letters so the gradient resolves smoothly regardless of casing/match length.
-const RAINBOW_STOPS = 14;
-
-let cachedMode: string | undefined;
-let cachedPalette: readonly string[] | undefined;
-
-/** Rainbow foreground escapes for the active color mode, compiled once per mode. */
-function rainbowPalette(): readonly string[] {
-	const mode = theme.getColorMode();
-	if (cachedPalette && cachedMode === mode) return cachedPalette;
-	const format = mode === "truecolor" ? "ansi-16m" : "ansi-256";
-	const palette: string[] = [];
-	for (let i = 0; i < RAINBOW_STOPS; i++) {
-		// Sweep red→violet (0..330°), stopping short of the wrap back to red.
-		const hue = Math.round((i / RAINBOW_STOPS) * 330);
-		palette.push(Bun.color(`hsl(${hue}, 90%, 62%)`, format) ?? "");
-	}
-	cachedMode = mode;
-	cachedPalette = palette;
-	return palette;
-}
-
-/** Paint each character of `word` with the next rainbow stop, resetting fg after. */
-function rainbow(word: string): string {
-	const palette = rainbowPalette();
-	const n = word.length;
-	let out = "";
-	let prev = "";
-	for (let i = 0; i < n; i++) {
-		const color = palette[Math.floor((i / n) * palette.length)] ?? palette[0] ?? "";
-		// Coalesce consecutive characters that resolve to the same stop.
-		if (color !== prev) {
-			out += color;
-			prev = color;
-		}
-		out += word[i];
-	}
-	return `${out}${FG_RESET}`;
-}
-
 /**
  * Rainbow-highlight every standalone "ultrathink" in `text` for editor display.
- * Adds only zero-width SGR escapes — the visible width is unchanged — and returns
- * the input untouched when the keyword is absent.
+ * Sweeps red→violet (hue 0..330), stopping short of the wrap back to red so the
+ * gradient resolves smoothly regardless of casing or match length.
  */
-export function highlightUltrathink(text: string): string {
-	if (!ULTRATHINK_PROBE.test(text)) return text;
-	return text.replace(ULTRATHINK_HIGHLIGHT, rainbow);
-}
+export const highlightUltrathink: (text: string) => string = createGradientHighlighter({
+	probe: /ultrathink/i,
+	highlight: /\bultrathink\b/gi,
+	stops: 14,
+	hue: t => t * 330,
+});

package/src/modes/utils/keybinding-matchers.ts

@@ -16,10 +16,21 @@ export function matchesAppInterrupt(data: string): boolean {
 	return matchesKey(data, "escape") || matchesKey(data, "esc");
 }
 
+/** Match the generic selector cancel keybinding. */
 export function matchesSelectCancel(data: string): boolean {
 	return getKeybindings().matches(data, "tui.select.cancel");
 }
 
+/** Match the generic selector up-navigation keybinding. */
+export function matchesSelectUp(data: string): boolean {
+	return getKeybindings().matches(data, "tui.select.up");
+}
+
+/** Match the generic selector down-navigation keybinding. */
+export function matchesSelectDown(data: string): boolean {
+	return getKeybindings().matches(data, "tui.select.down");
+}
+
 export function matchesAppExternalEditor(data: string): boolean {
 	const keybindings = getKeybindings();
 	const externalEditorKeys = keybindings.getKeys("app.editor.external");

package/src/prompts/system/memory-consolidation-system.md

@@ -0,0 +1,8 @@
+Summarize the memories below into 1-3 concise sentences.
+
+Preserve every fact, name, number, version, date, and decision exactly. Merge duplicates and near-duplicates; never repeat the same point. When memories conflict, state only the most recent as current. Do not invent, infer, or add anything that is not present in the memories. Output only the summary sentences, nothing else.
+
+Memories:
+{memories}
+
+Summary:

package/src/prompts/system/memory-extraction-system.md

@@ -0,0 +1,26 @@
+Extract durable, long-term memory items from the user message below.
+
+Output ONE item per line as a short plain-text statement: no JSON, no bullets, no numbering, no field labels.
+Capture only persistent, reusable information:
+- facts (name, role, employer, config, ports, versions, numbers)
+- explicit instructions to the assistant
+- stable preferences
+- dated events or deadlines
+
+Keep names, numbers, versions, and dates exact, in the message's original language. When a value is updated, output only the latest value. Ignore greetings, acknowledgements, small talk, weather, and one-off remarks.
+If nothing qualifies, output exactly: NO_FACTS
+
+Example
+Message: My name is Sam, I work at Globex, and I always use 2-space indents.
+Items:
+name is Sam
+works at Globex
+prefers 2-space indents
+
+Example
+Message: lol nice weather today, might grab a coffee later
+Items:
+NO_FACTS
+
+Message: {text}
+Items:

package/src/prompts/{commands/orchestrate.md → system/orchestrate-notice.md}

@@ -1,17 +1,5 @@
----
-name: orchestrate
-description: Drive a multi-phase task to completion via parallel subagents
----
-
-# Task
-
-$@
-
----
-
-# Orchestration Contract
-
-You are the **orchestrator** for the task above. Read it once, then execute under the rules below. The contract overrides any default tendency to yield early, narrate, or do work yourself.
+<system-notice>
+The user's message above is an **orchestration request**. Execute it as the orchestrator under the contract below. This contract overrides any default tendency to yield early, narrate, or do the work yourself.
 
 <role>
 You decompose, dispatch, verify, and iterate. You do **not** edit code. Every file mutation goes through a `task` subagent. Your tool budget is: reading for planning, `task` for dispatch, verification (`bun check`, `bun test`, `recipe`, `lsp diagnostics`), git via `bash`, and `todo_write` for tracking.
@@ -19,11 +7,11 @@ You decompose, dispatch, verify, and iterate. You do **not** edit code. Every fi
 
 <rules>
 1. **Do not yield until everything is closed.** A phase finishing is *not* a yield point — launch the next phase in the same turn. Stop only when every requested item is verifiably done, or you hit a concrete [blocked] state that genuinely requires the user.
-2. **Enumerate the full surface before dispatching.** If the task references audits, plans, checklists, phase lists, or file lists, expand them into a flat set of items in `todo_write`. "Most of them" or "the important ones" is failure. Re-read the source documents — do not work from memory.
+2. **Enumerate the full surface before dispatching.** If the request references audits, plans, checklists, phase lists, or file lists, expand them into a flat set of items in `todo_write`. "Most of them" or "the important ones" is failure. Re-read the source documents — do not work from memory.
 3. **Parallelize maximally.** Every set of edits with disjoint file scope MUST ship as one `task` batch. Serialize only when one subagent produces a contract (types, schema, shared module) the next consumes — and state the dependency when you do.
 4. **Each `task` assignment is self-contained.** Subagents have no shared context. Spell out: target files (≤3–5 explicit paths, no globs), the change with APIs and patterns, edge cases, and observable acceptance criteria. Do not assume they read the same plan you did.
 5. **Verify after every phase before launching the next.** Run the appropriate gate: `bun check` for types, package-scoped `bun test` for behavior, `lsp diagnostics` for changed files. If a phase introduced breakage, dispatch fix-up subagents *before* moving on. Never declare a phase done on a red tree.
-6. **Commit policy.** If the task asks for commits or the repo workflow expects them, commit after each green phase with a focused message. Never commit a red tree. Never commit work the user did not ask to commit.
+6. **Commit policy.** If the request asks for commits or the repo workflow expects them, commit after each green phase with a focused message. Never commit a red tree. Never commit work the user did not ask to commit.
 7. **Respawn, do not absorb.** If a subagent returns incomplete or wrong work, spawn a corrective subagent with the specific gap — do not silently fix it yourself.
 8. **No scope creep, no scope shrink.** Do not add work the user did not ask for. Do not relabel unfinished items as "follow-up", "v1", or "MVP" to imply completion.
 9. **Subagents do not verify, lint, or format.** Every `task` assignment MUST instruct the subagent to skip all gates and formatters. Their job is the edit only. You — the orchestrator — run verification and formatting **once** at the end of the phase across the union of changed files. Avoids redundant runs and racing formatter passes.
@@ -47,3 +35,4 @@ You decompose, dispatch, verify, and iterate. You do **not** edit code. Every fi
 - Marking todos done based on subagent self-reports without verifying the gate.
 - Summarizing progress in chat instead of advancing to the next phase.
 </anti-patterns>
+</system-notice>

package/src/prompts/system/system-prompt.md

@@ -54,7 +54,9 @@ With most FS/bash-like tools, static references to them will automatically resol
 - `skill://<name>`: Skill instructions
    - `/<path>`: File within a skill
 - `rule://<name>`: Rule details
+{{#if hasMemoryRoot}}
 - `memory://root`: Project memory summary
+{{/if}}
 - `agent://<id>`: Full agent output artifact
    - `/<path>`: JSON field extraction
 - `artifact://<id>`: Artifact content

package/src/prompts/system/tiny-title-system.md

@@ -0,0 +1,8 @@
+You generate concise terminal session titles.
+
+Input is one user message inside `<user-message>` tags.
+
+Return one specific 3-6 word title.
+Continue the assistant response after `<title>` and close it with `</title>`.
+
+NEVER include quotes, punctuation, markdown, commentary, or a second line.

package/src/prompts/tools/eval.md

@@ -44,6 +44,8 @@ output(*ids, format?="raw", query?=None, offset?=None, limit?=None) → str | di
     Read task/agent output by ID. Single id returns text/dict; multiple ids return a list.
 tool.<name>(args) → unknown
     Invoke any session tool by name. `args` is the tool's parameter object.
+llm(prompt, model?="default", system?=None, schema?=None) → str | dict
+    Oneshot, stateless LLM call (no history, no tools). `model` picks a tier: "smol" (fast), "default" (this session's model), "slow" (most capable). Pass `system` for a system prompt. Pass a JSON-Schema `schema` to force structured output and get the parsed object back; otherwise returns the completion text.
 ```
 </prelude>
 

package/src/prompts/tools/memory-edit.md

@@ -0,0 +1,8 @@
+Edit Mnemosyne long-term memories by id.
+
+Use only with ids returned by the `recall` tool. Operations:
+- `update`: replace content and/or importance for a working memory.
+- `forget`: permanently delete a working memory.
+- `invalidate`: softly supersede a working or episodic memory, optionally pointing at `replacement_id`.
+
+Prefer `invalidate` when a memory became stale but its history may still be useful. Use `forget` only for content that should be hard-deleted.

package/src/prompts/tools/task.md

@@ -30,15 +30,12 @@ Subagents have no conversation history. Every fact, file path, and direction the
 </parameters>
 
 <rules>
+- **Maximize batch width.** Spawn the widest parallel set the work decomposes into. NEVER spawn a single-task batch for divisible work, or defer work that could have been concurrent.
 - NEVER assign tasks to run project-wide build/test/lint. Caller verifies after the batch.
 - **Subagents do not verify, lint, or format.** Every assignment MUST instruct the subagent to skip all gates and formatters. You run them once at the end across the union of changed files — avoids redundant runs and racing formatter passes.
-{{#if ircEnabled}}
-- Each task: ≤3–5 explicit files. Overlapping file sets are tolerable when peers can coordinate via `irc`, but still fan out to a cluster when the scopes are cleanly separable.
-- No globs, no "update all", no package-wide scope.
-{{else}}
-- Each task: ≤3–5 explicit files. No globs, no "update all", no package-wide scope. Fan out to a cluster instead.
-{{/if}}
-- Pass large payloads via `local://<path>` URIs, not inline.
+- No globs, no "update all", no package-wide scope. Fan out.
+- Do not concern yourself with how agents might overlap on certain actions. Never use it as an excuse to go slower: they can resolve collisions in real-time with the harness facilities.
+- Pass large payloads via `local://<path>` URIs, not inline. {{#if contextEnabled}} (other than the context){{/if}}
 {{#if contextEnabled}}- Put shared constraints in `context` once; do not duplicate across assignments.{{/if}}
 - Prefer agents that investigate **and** edit in one pass; only spin a read-only discovery step when affected files are genuinely unknown.
 </rules>

package/src/sdk.ts

@@ -86,8 +86,8 @@ import type { HindsightSessionState } from "./hindsight/state";
 import { LocalProtocolHandler, type LocalProtocolOptions } from "./internal-urls";
 import { LSP_STARTUP_EVENT_CHANNEL, type LspStartupEvent } from "./lsp/startup-events";
 import { discoverAndLoadMCPTools, MCPManager, type MCPToolsLoadResult } from "./mcp";
-
 import { resolveMemoryBackend } from "./memory-backend";
+import { getMnemosyneSessionState, type MnemosyneSessionState } from "./mnemosyne/state";
 import asyncResultTemplate from "./prompts/tools/async-result.md" with { type: "text" };
 import { AgentRegistry, MAIN_AGENT_ID } from "./registry/agent-registry";
 import {
@@ -313,6 +313,8 @@ export interface CreateAgentSessionOptions {
 	taskDepth?: number;
 	/** Parent Hindsight state to alias for subagent memory tools. */
 	parentHindsightSessionState?: HindsightSessionState;
+	/** Parent Mnemosyne state to alias for subagent memory tools. */
+	parentMnemosyneSessionState?: MnemosyneSessionState;
 	/** Pre-allocated agent identity for IRC routing. Default: "0-Main" for top-level, parentTaskPrefix-derived for sub. */
 	agentId?: string;
 	/** Display name for the agent in IRC. Default: "main" or "sub". */
@@ -1187,6 +1189,7 @@ export async function createAgentSession(options: CreateAgentSessionOptions = {}
 				session ? session.trackEvalExecution(execution, abortController) : execution,
 			getSessionId: () => sessionManager.getSessionId?.() ?? null,
 			getHindsightSessionState: () => session?.getHindsightSessionState(),
+			getMnemosyneSessionState: () => getMnemosyneSessionState(session),
 			getAgentId: () => resolvedAgentId,
 			getToolByName: name => session?.getToolByName(name),
 			agentRegistry,
@@ -1593,11 +1596,8 @@ export async function createAgentSession(options: CreateAgentSessionOptions = {}
 			const promptTools = buildSystemPromptToolMetadata(tools, {
 				search_tool_bm25: { description: renderSearchToolBm25Description(discoverableToolsForDesc) },
 			});
-			const memoryInstructions = await resolveMemoryBackend(settings).buildDeveloperInstructions(
-				agentDir,
-				settings,
-				session,
-			);
+			const memoryBackend = resolveMemoryBackend(settings);
+			const memoryInstructions = await memoryBackend.buildDeveloperInstructions(agentDir, settings, session);
 
 			// Build combined append prompt: memory instructions + MCP server instructions
 			const serverInstructions = mcpManager?.getServerInstructions();
@@ -1634,6 +1634,7 @@ export async function createAgentSession(options: CreateAgentSessionOptions = {}
 				eagerTasks,
 				secretsEnabled,
 				workspaceTree: workspaceTreePromise,
+				memoryRootEnabled: memoryBackend.id === "local",
 			});
 
 			if (options.systemPrompt === undefined) {
@@ -2123,6 +2124,7 @@ export async function createAgentSession(options: CreateAgentSessionOptions = {}
 					agentDir,
 					taskDepth,
 					parentHindsightSessionState: options.parentHindsightSessionState,
+					parentMnemosyneSessionState: options.parentMnemosyneSessionState,
 				}),
 			),
 		);