oh-my-opencode 4.19.4 → 5.0.0-beta.10

package/packages/omo-codex/plugin/test/sync-skills-codex-compatibility.test.mjs

@@ -0,0 +1,101 @@
+import assert from "node:assert/strict";
+import { readdir, readFile } from "node:fs/promises";
+import { dirname, join } from "node:path";
+import test from "node:test";
+import { fileURLToPath } from "node:url";
+
+import {
+	codexHarnessToolCompatibility,
+	insertCodexCompatibilityGuidance,
+} from "../scripts/sync-skills.mjs";
+
+const frontmatter = "---\nname: fixture-sentinel\n---\n\n";
+const opencodeExample = "# SENTINEL_SECTION\ntask(SENTINEL_INPUT)\n";
+const pluginRoot = dirname(dirname(fileURLToPath(import.meta.url)));
+const repositoryRoot = join(pluginRoot, "..", "..");
+const opencodeToolPattern = /\b(?:call_omo_agent|background_output|team_[a-z_]+|task)\s*\(/;
+
+test("#given sentinel OpenCode tool content #when compatibility guidance is inserted #then the production artifact is placed after frontmatter and before the tool token", () => {
+	const input = `${frontmatter}${opencodeExample}`;
+
+	const actual = insertCodexCompatibilityGuidance(input);
+
+	assert.equal(actual, `${frontmatter}${codexHarnessToolCompatibility}${opencodeExample}`);
+	assert.ok(actual.indexOf(codexHarnessToolCompatibility) < actual.indexOf(opencodeExample));
+});
+
+test("#given already transformed sentinel content #when compatibility guidance is inserted again #then the transform is idempotent", () => {
+	const once = insertCodexCompatibilityGuidance(`${frontmatter}${opencodeExample}`);
+
+	const twice = insertCodexCompatibilityGuidance(once);
+
+	assert.equal(twice, once);
+});
+
+test("#given a stale generated compatibility block #when guidance is inserted #then the production artifact replaces it", () => {
+	const staleSentinel = "STALE_GENERATED_SENTINEL";
+	const staleGuidance = codexHarnessToolCompatibility.replace("multi_agent_v1.spawn_agent", staleSentinel);
+	assert.notEqual(staleGuidance, codexHarnessToolCompatibility);
+
+	const actual = insertCodexCompatibilityGuidance(`${frontmatter}${staleGuidance}${opencodeExample}`);
+
+	assert.equal(actual, `${frontmatter}${codexHarnessToolCompatibility}${opencodeExample}`);
+	assert.equal(actual.includes(staleSentinel), false);
+});
+
+test("#given a custom compatibility block before an OpenCode tool token #when guidance is inserted #then the custom block is preserved", () => {
+	const generatedHeading = codexHarnessToolCompatibility.slice(0, codexHarnessToolCompatibility.indexOf("\n") + 1);
+	const customBlock = `${generatedHeading}\nCUSTOM_BLOCK_SENTINEL\n\n`;
+	const input = `${frontmatter}${customBlock}${opencodeExample}`;
+
+	const actual = insertCodexCompatibilityGuidance(input);
+
+	assert.equal(actual, input);
+});
+
+test("#given an exported template wrapper containing an OpenCode tool token #when guidance is inserted #then wrapper bytes are preserved", () => {
+	const templateWrapper = "export const TEMPLATE_SENTINEL = `task(INPUT_SENTINEL)`;\n";
+
+	const actual = insertCodexCompatibilityGuidance(templateWrapper);
+
+	assert.equal(actual, `${codexHarnessToolCompatibility}${templateWrapper}`);
+});
+
+test("#given real shared skills that need Codex translation #when aggregate skills are synced #then each generated skill injects the production compatibility artifact once before any remaining OpenCode tool token", async () => {
+	const sharedSkillsRoot = join(repositoryRoot, "shared-skills", "skills");
+	const entries = await readdir(sharedSkillsRoot, { withFileTypes: true });
+	const syncScript = await readFile(join(pluginRoot, "scripts", "sync-skills.mjs"), "utf8");
+	const componentSkillNames = new Set(
+		[...syncScript.matchAll(/\[\s*"([^"]+)"\s*,\s*"components\//g)].map((match) => match[1]),
+	);
+	let checked = 0;
+
+	for (const entry of entries) {
+		if (!entry.isDirectory()) continue;
+		if (componentSkillNames.has(entry.name)) continue;
+		const source = await readFile(join(sharedSkillsRoot, entry.name, "SKILL.md"), "utf8");
+		const sourceToolToken = source.match(opencodeToolPattern)?.[0];
+		if (!sourceToolToken) continue;
+		const generated = await readFile(join(pluginRoot, "skills", entry.name, "SKILL.md"), "utf8");
+
+		assert.equal(generated.split(codexHarnessToolCompatibility).length - 1, 1, entry.name);
+		const compatibilityIndex = generated.indexOf(codexHarnessToolCompatibility);
+		assert.ok(compatibilityIndex >= 0, entry.name);
+		const firstGeneratedToolIndex = generated.search(opencodeToolPattern);
+		assert.ok(firstGeneratedToolIndex >= compatibilityIndex, entry.name);
+		assert.ok(firstGeneratedToolIndex < compatibilityIndex + codexHarnessToolCompatibility.length, entry.name);
+		checked += 1;
+	}
+
+	assert.ok(checked > 0, "at least one shared skill must exercise compatibility injection");
+});
+
+test("#given the aggregate sync implementation #when its skill adaptation pipeline is inspected #then it still applies compatibility guidance before overlays", async () => {
+	const script = await readFile(join(pluginRoot, "scripts", "sync-skills.mjs"), "utf8");
+
+	assert.match(
+		script,
+		/applyCodexSkillOverlays\(\s*skillName,\s*insertCodexCompatibilityGuidance\(content\),?\s*\)/,
+	);
+	assert.match(script, /await adaptSkillForCodex\(skillName\)/);
+});

package/packages/omo-codex/plugin/test/sync-skills-test-support.mjs

@@ -90,9 +90,9 @@ const startWorkCodexCompletion = `When all top-level checkboxes in \`## TODOs\`
 4. Remove or mark the Boulder work as completed.
 5. Print an \`ORCHESTRATION COMPLETE\` block with the plan path, verification commands, Global Review and Debugging Gate verdict, artifacts, and cleanup receipts.`;
 
-const startWorkOriginalHardRule = "- No completion claim while an applicable ultraqa adversarial class was never probed. Each applicable class needs a captured observable result; each skipped class needs a one-line not-applicable reason in the ledger.\n- No PR/branch implementation, review, or merge in the main worktree; use the task-owned git worktree.\n- No unprefixed session ids in Boulder state. Codex sessions are always `codex:<session_id>`.";
+const startWorkOriginalHardRule = "- No completion claim while an applicable ultraqa adversarial class was never probed. Each applicable class needs a captured observable result; each skipped class needs a one-line not-applicable reason in the ledger.\n- No PR/branch implementation, review, or merge in the main worktree; use the task-owned git worktree.\n- No unprefixed session ids in Boulder state. Sessions are always recorded as `codex:<session_id>`.";
 
-const startWorkCodexHardRule = "- No completion claim while an applicable ultraqa adversarial class was never probed. Each applicable class needs a captured observable result; each skipped class needs a one-line not-applicable reason in the ledger.\n- No `ORCHESTRATION COMPLETE`, final response, PR creation, PR handoff, or merge before the Global Review and Debugging Gate passes with recorded evidence.\n- No PR/branch implementation or review in the main worktree; create or use a task-owned git worktree first.\n- No unprefixed session ids in Boulder state. Codex sessions are always `codex:<session_id>`.";
+const startWorkCodexHardRule = "- No completion claim while an applicable ultraqa adversarial class was never probed. Each applicable class needs a captured observable result; each skipped class needs a one-line not-applicable reason in the ledger.\n- No `ORCHESTRATION COMPLETE`, final response, PR creation, PR handoff, or merge before the Global Review and Debugging Gate passes with recorded evidence.\n- No PR/branch implementation or review in the main worktree; create or use a task-owned git worktree first.\n- No unprefixed session ids in Boulder state. Sessions are always recorded as `codex:<session_id>`.";
 
 const reviewWorkCodexGate = `
 When \`review-work\` is used as a final implementation, PR, or \`$start-work\`
@@ -127,8 +127,8 @@ prefixes when identity is needed.
 `;
 const reviewWorkCodexGatePattern = new RegExp(reviewWorkCodexGate.replace(/[.*+?^${}()|[\]\\]/g, "\\$&"));
 
-const ulwResearchOriginalDeliveryGates = "### The delivery gates \u2014 every gate must PASS, in order\n\nNothing reaches the user until the gates pass:\n\n1. **Visual QA (always).** Render the produced artifact back to images \u2014 PDF pages to PNG, the HTML in a real browser \u2014 and look at them: missing or broken figures, clipped tables, overflowing CJK text, blank pages, unreadable chart labels, wrong page breaks. Fix and re-render until the pages are clean. Reading the source markup is not visual QA; inspect the pixels.\n2. **Proofread gate \u2014 `task(category=\"writing\", ...)`.** Hand the final text to a dedicated `writing` worker whose only job is language: grammar, spelling, punctuation, terminology consistency, and whether the prose reads NATIVELY in the report's own language. It returns a defect list; fix every item and re-run the gate on the delta. Deliver only on a clean pass \u2014 this gate runs BEFORE the first delivery, not after the user finds the typo.";
-const ulwResearchCodexDeliveryGate = "### The delivery gate \u2014 visual QA must PASS\n\nNothing reaches the user until the gate passes:\n\n**Visual QA (always).** Render the produced artifact back to images \u2014 PDF pages to PNG, the HTML in a real browser \u2014 and look at them: missing or broken figures, clipped tables, overflowing CJK text, blank pages, unreadable chart labels, wrong page breaks. Fix and re-render until the pages are clean. Reading the source markup is not visual QA; inspect the pixels.";
+const ulwResearchOriginalDeliveryGates = "### The delivery gates \u2014 every gate must PASS, in order\n\nNothing reaches the user until the gates pass:\n\n1. **Visual QA (always).** Render the produced artifact back to images \u2014 PDF pages to PNG, the HTML in a real browser \u2014 and look at them: missing or broken figures, images stretched or spilling their containers, diagram or chart text rendered off the spec's font or palette, clipped tables, overflowing CJK text, blank pages, unlabeled chart values, wrong page breaks. Fix and re-render until the pages are clean. Reading the source markup is not visual QA; inspect the pixels.\n2. **Proofread gate \u2014 `task(category=\"writing\", ...)`.** Hand the final text to a dedicated `writing` worker whose only job is language: grammar, spelling, punctuation, terminology consistency, and whether the prose reads NATIVELY in the report's own language. It returns a defect list; fix every item and re-run the gate on the delta. Deliver only on a clean pass \u2014 this gate runs BEFORE the first delivery, not after the user finds the typo.";
+const ulwResearchCodexDeliveryGate = "### The delivery gate \u2014 visual QA must PASS\n\nNothing reaches the user until the gate passes:\n\n**Visual QA (always).** Render the produced artifact back to images \u2014 PDF pages to PNG, the HTML in a real browser \u2014 and look at them: missing or broken figures, images stretched or spilling their containers, diagram or chart text rendered off the spec's font or palette, clipped tables, overflowing CJK text, blank pages, unlabeled chart values, wrong page breaks. Fix and re-render until the pages are clean. Reading the source markup is not visual QA; inspect the pixels.";
 
 export function removeCodexSkillOverlays(skillName, content) {
 	if (skillName === "ulw-research") {

package/packages/omo-codex/plugin/test/sync-skills.test.mjs

@@ -5,8 +5,6 @@ import test from "node:test";
 import { fileURLToPath } from "node:url";
 import { sharedSkillsRootPath } from "@oh-my-opencode/shared-skills";
 import {
-	CONTEXT_PRESSURE_SKILL_BUDGET_BYTES,
-	assertPackagedContentMatches,
 	componentSkillSources,
 	expectedSkills,
 	listSkillFiles,
@@ -16,7 +14,6 @@ import {
 
 const root = dirname(dirname(fileURLToPath(import.meta.url)));
 const repoRoot = join(root, "..", "..", "..");
-const opencodeOnlyToolPattern = /\b(?:call_omo_agent|background_output|team_[a-z_]+|task)\s*\(/;
 const generatedSkillMetadataFiles = new Set(["agents/openai.yaml"]);
 
 async function readPackagedSkillFile(...segments) {
@@ -131,7 +128,6 @@ test("#given a shared skill name collides with a Codex component skill #when agg
 	// when / then
 	assert.notEqual(removeCodexCompatibilityGuidance(aggregateSkill), removeCodexCompatibilityGuidance(sharedSkill));
 	assert.equal(removeCodexCompatibilityGuidance(aggregateSkill), removeCodexCompatibilityGuidance(componentSkill));
-	assert.match(aggregateSkill, /multi_agent_v1/);
 });
 
 test("#given shared skill source tests #when aggregate Codex skills are synced #then source tests are not packaged", async () => {
@@ -191,9 +187,7 @@ test("#given synced ulw-loop skill #when Codex hint metadata is inspected #then
 	// then
 	assert.match(skill, /^---\r?\nname: ulw-loop\r?\n/m);
 	assert.match(interfaceMetadata, /display_name: "\(OmO\) ulw-loop"/);
-	assert.doesNotMatch(interfaceMetadata, /ulw-loop \/ ulw-loop/);
-	assert.match(interfaceMetadata, /short_description: "Goal-like ultrawork loop for systematic decomposition"/);
-	assert.match(interfaceMetadata, /default_prompt: "Use \$ulw-loop/);
+	assert.match(interfaceMetadata, /^\s*default_prompt:\s*".+"$/m);
 });
 
 test("#given synced ulw-loop skill #when Codex hint metadata is inspected #then ulw-loop remains discoverable as an alias", async () => {
@@ -229,92 +223,11 @@ test("#given synced git-master skill #when inspected #then commits and git histo
 
 	// then
 	assert.match(skill, /^---\r?\nname: git-master\r?\n/m);
-	assert.match(skill, /MUST USE whenever a task needs a commit or git-history investigation/);
-	assert.match(skill, /Commit only the user's requested changes/);
-	assert.match(skill, /Choose the Git tool by the question/);
-	assert.match(skill, /git log -S "text"/);
-	assert.match(skill, /git blame -L start,end -- file/);
 	assert.match(interfaceMetadata, /display_name: "\(OmO\) git-master"/);
 	assert.match(interfaceMetadata, /- "git commit"/);
 	assert.match(interfaceMetadata, /- "history search"/);
 });
 
-test("#given synced ulw-loop skill #when worker guidance is inspected #then context-hygiene guidance matches the source", async () => {
-	// given
-	const sourceSkill = await readFile(
-		join(root, "components", "ulw-loop", "skills", "ulw-loop", "references", "full-workflow.md"),
-		"utf8",
-	);
-	const syncedSkill = await readFile(join(root, "skills", "ulw-loop", "SKILL.md"), "utf8");
-	const syncedWorkflow = await readFile(join(root, "skills", "ulw-loop", "references", "full-workflow.md"), "utf8");
-	// ulw-loop is V2-primary (gpt-5.6 sol/terra use the flat `wait_agent`); the `multi_agent_v1.*`
-	// namespace is documented only as the v1 fallback, so the wait_agent refs accept the bare token.
-	const requiredPatterns = [
-		["wait_agent ref", /\bwait_agent\b/],
-		["local spawned-name tracking", /Track spawned agent names locally/],
-		["wait_agent mailbox path", /wait_agent.*mailbox signals/],
-		["progress status contract", /WORKING:/],
-		["long-running plan/reviewer background guidance", /Plan and reviewer agents may run for a long time/],
-		["bounded plan/reviewer polling", /wait_agent.*cycles/],
-		["exponential backoff wait guard", /double the timeout up to ~5 minutes/],
-		["git-master checkpointing", /git-master/],
-		["touched-path commit-style probe", /touched-path commit history/],
-		["verified work-unit commit", /verified work unit/],
-		["observed commit style", /commit in the observed style/],
-	];
-
-	// when / then
-	for (const [label, pattern] of requiredPatterns) {
-		assert.match(sourceSkill, pattern, `source skill missing ${label}`);
-		assert.match(syncedWorkflow, pattern, `synced workflow missing ${label}`);
-	}
-	assert.match(syncedSkill, /references\/full-workflow\.md/);
-	assert.match(syncedSkill, /wait_agent/);
-	assert.match(syncedSkill, /close_agent/);
-});
-
-test("#given packaged start-work skill #when inspected #then no-plan bootstrap and adversarial verification contracts are shipped", async () => {
-	// given
-	const skillFile = await readPackagedSkillFile("start-work", "SKILL.md");
-
-	// when / then
-	assertPackagedContentMatches(skillFile, [
-		["executes Prometheus plan with Boulder state", /Prometheus work plan[\s\S]*Boulder state/],
-		["bootstraps ulw-plan when no selectable plan exists", /no selectable plan[\s\S]*ulw-plan|ulw-plan[\s\S]*no selectable plan/i],
-		["does not execute work without an approved plan", /approved plan[\s\S]*(?:before|prior to)[\s\S]*execution|execution[\s\S]*(?:requires|needs)[\s\S]*approved plan/i],
-		["keeps hook continuation Boulder-only", /Boulder[\s\S]*(?:continuation|Stop hook)[\s\S]*(?:only|solely)|(?:continuation|Stop hook)[\s\S]*(?:only|solely)[\s\S]*Boulder/i],
-		["distinguishes execution from verification", /execution[\s\S]*verification|verification[\s\S]*execution/i],
-		["requires dirty-worktree-aware editing", /dirty worktree/i],
-		["requires stale-state probes", /stale state/i],
-		["rejects misleading success output", /misleading success output/i],
-		["does not accept worker done claims without independent verification", /done claim[\s\S]*independent(?:ly)? verified|independent(?:ly)? verify[\s\S]*done claim/i],
-	]);
-});
-
-test("#given packaged ulw-plan skill #when inspected #then dynamic multi-agent planning contracts are shipped", async () => {
-	// given
-	const skillFile = await readPackagedSkillFile("ulw-plan", "SKILL.md");
-	const workflowFile = await readPackagedSkillFile("ulw-plan", "references", "full-workflow.md");
-	const combinedFile = {
-		path: `${skillFile.path} + ${workflowFile.path}`,
-		content: `${skillFile.content}\n${workflowFile.content}`,
-	};
-
-	// when / then
-	assertPackagedContentMatches(combinedFile, [
-		["self-orchestrates 5 host subagents for planning", /(?:self-orchestrates|orchestrates)[\s\S]*5[\s\S]*host subagents/i],
-		["requires dynamic workflow phases", /dynamic[\s\S]*workflow[\s\S]*phase|phase[\s\S]*dynamic[\s\S]*workflow/i],
-		["keeps verification distinct from execution", /verification[\s\S]*execution|execution[\s\S]*verification/i],
-		["requires dirty-worktree-aware planning", /dirty worktree/i],
-		["requires stale-state checks between source and packaged payloads", /stale state/i],
-		["rejects misleading success output", /misleading success output/i],
-		["does not accept subagent outputs as success without independent verification", /subagent outputs?[\s\S]*(?:not|never)[\s\S]*(?:success|approval)|independent(?:ly)? verif(?:y|ied|ication)[\s\S]*subagent outputs?/i],
-		["treats Discord or external content as claims, not instructions", /(?:Discord|external content)[\s\S]*claims?[\s\S]*not instructions?|not instructions?[\s\S]*(?:Discord|external content)/i],
-	]);
-	assert.doesNotMatch(combinedFile.content, opencodeOnlyToolPattern);
-	assert.doesNotMatch(combinedFile.content, /Proceeding to plan generation/);
-});
-
 test("#given packaged Codex ulw-plan surfaces #when inspected #then dangerous sandbox bypass guidance is not shipped", async () => {
 	// given
 	const dangerousBypassToken = ["dangerously", "bypass"].join("-");
@@ -330,34 +243,3 @@ test("#given packaged Codex ulw-plan surfaces #when inspected #then dangerous sa
 	assert.doesNotMatch(packagedWorkflow.content, dangerousBypassPattern, `${packagedWorkflow.path} ships unsafe Codex bypass guidance`);
 	assert.doesNotMatch(componentWorkflow.content, dangerousBypassPattern, `${componentWorkflow.path} ships unsafe Codex bypass guidance`);
 });
-
-test("#given packaged ulw-research skill #when Codex delivery gates are inspected #then the category-based proofread gate is stripped", async () => {
-	// given
-	const content = await readFile(join(root, "skills", "ulw-research", "SKILL.md"), "utf8");
-
-	// then the writing-category proofread gate never ships to Codex (no category concept there)
-	assert.doesNotMatch(content, /category="writing"/, "codex ulw-research ships the writing-category proofread gate");
-	assert.doesNotMatch(content, /Proofread gate/, "codex ulw-research still references the proofread gate");
-	// and the visual-QA delivery gate survives as the single gate
-	assert.match(content, /### The delivery gate \u2014 visual QA must PASS/u, "codex ulw-research lost the visual-QA delivery gate");
-	assert.match(content, /Visual QA \(always\)/, "codex ulw-research lost the visual-QA gate body");
-});
-
-test("#given context-pressure-prone skills #when bundled for Codex #then the eagerly loaded payload stays budgeted", async () => {
-	// given
-	const skillsRoot = join(root, "skills");
-	const skillNames = ["debugging", "ulw-loop"];
-
-	// when
-	let totalBytes = 0;
-	for (const skillName of skillNames) {
-		const content = await readFile(join(skillsRoot, skillName, "SKILL.md"), "utf8");
-		totalBytes += Buffer.byteLength(content, "utf8");
-	}
-
-	// then
-	assert.ok(
-		totalBytes <= CONTEXT_PRESSURE_SKILL_BUDGET_BYTES,
-		`debugging + ulw-loop eager payload is ${totalBytes} bytes, above ${CONTEXT_PRESSURE_SKILL_BUDGET_BYTES}`,
-	);
-});

package/packages/omo-codex/plugin/test/teammode-archive-ambiguity.test.mjs

@@ -1,48 +1,8 @@
 import assert from "node:assert/strict";
-import { readFileSync } from "node:fs";
-import { join } from "node:path";
 import test from "node:test";
 
-import { root } from "./aggregate-plugin-fixture.mjs";
 import { cleanupTeamRoot, createTeamRoot, readTeamJson, runTeam } from "./teammode-safety-fixture.mjs";
 
-const teammodeSkillPath = join("components", "teammode", "skills", "teammode", "SKILL.md");
-
-test("#given Codex App thread ids can be ambiguous across hosts #when archive guidance is read #then it separates app-thread archival from team-state archival", () => {
-	const body = readFileSync(join(root, teammodeSkillPath), "utf8");
-
-	assert.match(
-		body,
-		/Ambiguous Codex thread id|ambiguous across hosts/i,
-		"archive guidance must name the Codex App multi-host ambiguity failure instead of only saying the archive tool is unavailable",
-	);
-	assert.match(
-		body,
-		/app-thread archival blocker/i,
-		"archive guidance must classify ambiguous app-thread ids as app-thread archival blockers",
-	);
-	assert.match(
-		body,
-		/record .*team log/i,
-		"archive guidance must require a durable team-log record when app-thread archival cannot be proven",
-	);
-	assert.match(
-		body,
-		/continue .*team-state archive/i,
-		"archive guidance must let the team-state archive proceed after recording the app-thread archival blocker",
-	);
-	assert.match(
-		body,
-		/archive --note/i,
-		"archive guidance must route ambiguous app-thread archive blockers into archive --note evidence",
-	);
-	assert.match(
-		body,
-		/Do not delete/i,
-		"archive guidance must keep evidence available until ambiguous app-thread archival is resolved",
-	);
-});
-
 test("#given app-thread archival is blocked by an ambiguous id #when archive runs with a note #then team state is archived and keeps the blocker in the log", () => {
 	const tempRoot = createTeamRoot("omo-codex-teammode-archive-ambiguity-");
 	try {

package/packages/omo-codex/plugin/test/teammode-communication.test.mjs

@@ -3,14 +3,12 @@ import { readFileSync } from "node:fs";
 import { join } from "node:path";
 import test from "node:test";
 
-import { root } from "./aggregate-plugin-fixture.mjs";
 import { cleanupTeamRoot, createTeamRoot, runTeam, teamDir } from "./teammode-safety-fixture.mjs";
 
-// The teammode guide.md is what every member actually reads. The real run of leader session
-// 019eed1e- proved members default to silence: 10 members, ~6 min, 4 end-of-work reports, zero
-// peer messages, zero mid-task reports, zero send_message_to_thread calls - because the prose said
-// "over-communicate relentlessly" (aspirational, no trigger) and never named the tool or targets.
-// These tests pin the fix: a concrete push-communication protocol is the default the manual renders.
+// The teammode guide.md is what every member actually reads. A real leader-session run proved members
+// default to silence when the guide never names the tool or the address book, so the tests below pin
+// only the machine tokens the generator must surface: the cross-thread tool name and the team.json
+// field paths members address messages by. Guide/prompt wording itself is prose and stays unpinned.
 
 function buildTwoMemberTeam(tempRoot, sessionId) {
 	runTeam(tempRoot, "init", "--name", "Comms", "--session-name", "frequent", "--session", sessionId);
@@ -39,69 +37,15 @@ test("#given a bound team #when the member field manual renders #then it names t
 		cleanupTeamRoot(tempRoot);
 	}
 });
-
-test("#given a bound team #when the member field manual renders #then it gives concrete per-moment cadence triggers, not aspirational 'constantly'", () => {
-	const tempRoot = createTeamRoot("omo-codex-teammode-comms-");
-	try {
-		buildTwoMemberTeam(tempRoot, "comms-cadence");
-		const guide = readGuide(tempRoot, "comms-cadence");
-
-		// then - a concrete heartbeat cadence the model can bind to
-		assert.match(guide, /every few tool calls/, "guide must give a concrete heartbeat trigger");
-		assert.match(guide, /WORKING:/, "guide must keep the WORKING heartbeat marker");
-		assert.match(guide, /BLOCKED:/, "guide must keep the BLOCKED marker");
-		// then - an explicit peer-notify rule so cross-cutting findings are pushed, not siloed
-		assert.match(guide, /to a \*\*peer\*\*/, "guide must have an explicit peer-notify rule");
-		assert.match(guide, /touches their slice/, "guide must trigger peer messages on slice-touching findings");
-		// then - the reassurance WHY that ties the two halves together: frequent reporting is what earns
-		// uninterrupted focus, so the leader has no reason to interrupt a member who keeps it informed
-		assert.match(guide, /uninterrupted focus/i, "guide must tell members frequent updates buy uninterrupted focus from the leader");
-		// then - the end-only directive that previously won is gone (it licensed batch-at-the-end reporting)
-		assert.doesNotMatch(guide, /the moment your work is complete/, "the standalone end-only report directive must be removed");
-	} finally {
-		cleanupTeamRoot(tempRoot);
-	}
-});
-
-test("#given a member bootstrap trigger #when it is generated #then it points at the push tool and a continuous cadence, not an end-of-work report", () => {
+test("#given a member bootstrap trigger #when it is generated #then it names the cross-thread push tool", () => {
 	const tempRoot = createTeamRoot("omo-codex-teammode-comms-");
 	try {
 		buildTwoMemberTeam(tempRoot, "comms-bootstrap");
 		const prompt = runTeam(tempRoot, "member-prompt", "--team", "comms-bootstrap", "--id", "A").stdout;
 
-		// then - the first thing a member reads tells it to push via the tool as it works
+		// then - the first thing a member reads names the tool it can push through
 		assert.match(prompt, /codex_app\.send_message_to_thread/, "bootstrap must name the push tool");
-		assert.match(prompt, /never just one report at the end/, "bootstrap must reject end-only reporting");
-		// then - it no longer frames reporting as a one-shot completion event
-		assert.doesNotMatch(prompt, /report to the leader the moment you are done/, "the end-only bootstrap line must be removed");
 	} finally {
 		cleanupTeamRoot(tempRoot);
 	}
 });
-
-test("#given the leader-facing SKILL.md #when it describes communication #then it requires --session so members can reach the leader and frames members as pushing", () => {
-	const skill = readFileSync(join(root, "components", "teammode", "skills", "teammode", "SKILL.md"), "utf8");
-
-	// then - the leader is told members PUSH to it (not only that the leader sends)
-	assert.match(skill, /Members push to you/i, "SKILL.md must frame members as pushing to the leader");
-	// then - the --session requirement is stated WITH its reason (member->leader push needs a real leader thread)
-	assert.match(skill, /--session <your own thread id>/, "SKILL.md must require init --session <your own thread id>");
-	assert.match(skill, /stuck polling|cannot report to you/i, "SKILL.md must explain why --session matters");
-});
-
-// The leader (a GPT-5.5 main session) over-polls when nothing reframes silence as work: real jobdori
-// run 019f07a6- fired 6 leader->child pings, 4 pure "WORKING CHECK"/"NUDGE" nags that interrupted the
-// child mid-slice. The member guide already mandates frequent reporting; the missing half is telling the
-// leader to WAIT. These tests pin that the leader-facing skill reframes quiet-as-working and gates any
-// outbound ping behind concrete decision rules instead of letting anxiety drive a barrage.
-test("#given the leader-facing SKILL.md #when it covers waiting on members #then it reframes a quiet member as working and gates leader pings behind decision rules", () => {
-	const skill = readFileSync(join(root, "components", "teammode", "skills", "teammode", "SKILL.md"), "utf8");
-
-	// then - a quiet member between heartbeats is reframed as working, not a stall to chase
-	// (\s+ tolerates markdown line-wrapping between the two words)
-	assert.match(skill, /working,\s+not stalled/i, "SKILL.md must reframe a quiet member as working, not stalled");
-	// then - outbound leader messages are gated behind explicit decision rules, not anxiety
-	assert.match(skill, /Message a member only when/i, "SKILL.md must gate leader pings behind a concrete decision rule");
-	// then - the anti-pattern nag pings the real run produced are named as what NOT to send
-	assert.match(skill, /are you\s+done\?/i, "SKILL.md must name the 'are you done?' style nag as forbidden");
-});

package/packages/omo-codex/plugin/test/teammode-thread-links.test.mjs

@@ -71,46 +71,13 @@ test("#given a worktree-backed team has bound member threads #when guide and sta
 		const status = runTeam(tempRoot, "status", "--team", sessionId).stdout;
 		const prompt = runTeam(tempRoot, "member-prompt", "--team", sessionId, "--id", "A").stdout;
 
+		// then - guide, status, and bootstrap prompt all surface the member's bound thread link and cwd
 		assert.match(guide, /codex:\/\/threads\/019ef350-ee78-72a3-bd5e-e40cebc3d814/);
 		assert.match(guide, /codex:\/\/threads\/019ef2dd-5b99-7ae2-8461-ffa6ca309d23/);
 		assert.match(guide, /worktree `\/tmp\/review-worktree`/);
 		assert.match(status, /link=codex:\/\/threads\/019ef350-ee78-72a3-bd5e-e40cebc3d814/);
-		assert.match(prompt, /Your Codex thread link is codex:\/\/threads\/019ef350-ee78-72a3-bd5e-e40cebc3d814/);
-		assert.match(prompt, /Work inside `\/tmp\/review-worktree`/);
-		assert.match(prompt, /Before editing, verify that your assigned worktree exists and contains repo files/);
-		assert.match(prompt, /BLOCKED: worktree not ready/);
-	} finally {
-		cleanupTeamRoot(tempRoot);
-	}
-});
-
-test("#given a worktree member has no cwd bound yet #when guide and prompt render #then they tell Codex to wait for readiness", () => {
-	const tempRoot = createTeamRoot("omo-codex-teammode-worktree-wait-");
-	try {
-		const sessionId = "worktree-wait";
-		runTeam(tempRoot, "init", "--name", "Wait", "--session-name", "worktrees", "--session", sessionId, "--worktree", "--base-branch", "main");
-		addMember(tempRoot, sessionId, {
-			id: "A",
-			name: "worktree-driver",
-			focus: "drive a pending Codex worktree thread",
-			lens: "ownership",
-			deliverable: "ready worktree report",
-		});
-		addMember(tempRoot, sessionId, {
-			id: "B",
-			name: "integration-watch",
-			focus: "watch integration boundaries",
-			lens: "perspective",
-			deliverable: "integration notes",
-		});
-		runTeam(tempRoot, "bind-thread", "--team", sessionId, "--id", "A", "--thread", "019ef350-ee78-72a3-bd5e-e40cebc3d814");
-
-		const guide = readFileSync(`${teamDir(tempRoot, sessionId)}/guide.md`, "utf8");
-		const prompt = runTeam(tempRoot, "member-prompt", "--team", sessionId, "--id", "A").stdout;
-
-		assert.match(guide, /If Codex returns only `pendingWorktreeId`/);
-		assert.match(guide, /wait until Codex surfaces a real thread id/);
-		assert.match(prompt, /If this prompt arrived before your Codex worktree checkout is ready/);
+		assert.match(prompt, /codex:\/\/threads\/019ef350-ee78-72a3-bd5e-e40cebc3d814/);
+		assert.match(prompt, /\/tmp\/review-worktree/);
 	} finally {
 		cleanupTeamRoot(tempRoot);
 	}

package/packages/omo-codex/plugin/test/teammode-transport.test.mjs

@@ -3,7 +3,6 @@ import { existsSync, readFileSync, writeFileSync } from "node:fs";
 import { join } from "node:path";
 import test from "node:test";
 
-import { root } from "./aggregate-plugin-fixture.mjs";
 import {
 	cleanupTeamRoot,
 	createTeamRoot,
@@ -53,45 +52,6 @@ function addV2Members(tempRoot, sessionId) {
 	);
 }
 
-test("#given flat V2 tools are available #when the leader reads teammode #then transport selection and user announcement precede init", () => {
-	const skill = readFileSync(join(root, "components", "teammode", "skills", "teammode", "SKILL.md"), "utf8");
-	const inspectIndex = skill.indexOf("Inspect your active tool list");
-	const initIndex = skill.indexOf('team.mjs" init');
-
-	assert.notEqual(inspectIndex, -1, "skill must inspect the active tool list");
-	assert.notEqual(initIndex, -1, "skill must still document init");
-	assert.ok(inspectIndex < initIndex, "transport inspection must happen before init");
-	assert.match(skill, /spawn_agent.*task_name[\s\S]*send_message[\s\S]*followup_task[\s\S]*wait_agent[\s\S]*list_agents[\s\S]*interrupt_agent/);
-	assert.match(skill, /tell the user|announce.*transport/i);
-	assert.match(skill, /MultiAgentV2/i);
-	assert.match(skill, /Codex App.*fallback/i);
-});
-
-test("#given neither transport's tools are visible #when the leader reads teammode #then search hits are revalidated and fallback is capability-aware without team state", () => {
-	const skill = readFileSync(join(root, "components", "teammode", "skills", "teammode", "SKILL.md"), "utf8");
-
-	const searchIndex = skill.indexOf("tool_search");
-	const unavailableIndex = skill.indexOf("Teammode unavailable");
-	assert.notEqual(searchIndex, -1, "skill must route hidden tools through the tool_search check");
-	assert.notEqual(unavailableIndex, -1, "skill must give the leader unavailable announcement templates");
-	assert.ok(searchIndex < unavailableIndex, "tool_search must precede the unavailable conclusion");
-	assert.match(skill, /revalidate.*COMPLETE.*compatible transport set/is, "a deferred-tool hit must be revalidated as a complete transport");
-	assert.match(skill, /another visible plain-subagent mechanism.*spawn.*communicate.*observe/is, "plain-subagent fallback requires a complete visible lifecycle");
-	assert.match(skill, /Otherwise continue serially.*capability limitation/is, "missing all transports must continue serially rather than promise workers");
-	assert.match(skill, /Do NOT run `init`/, "the fallback must not create team state");
-	assert.doesNotMatch(skill, /splitting the work across plain fire-and-forget subagents/i, "the skill must not promise unavailable plain subagents");
-});
-
-test("#given a MultiAgentV2 team #when the leader spawns members #then guidance limits calls to exposed V2 fields", () => {
-	const skill = readFileSync(join(root, "components", "teammode", "skills", "teammode", "SKILL.md"), "utf8");
-
-	assert.match(skill, /using only the V2 schema fields/i);
-	assert.match(skill, /`task_name`[\s\S]*`message`[\s\S]*`fork_turns`/);
-	assert.match(skill, /role, priority, or task-specific routing instruction in `message`/i);
-	assert.match(skill, /V2 does not accept[\s\S]*`agent_type`, `model`, `reasoning_effort`, or `service_tier`/);
-	assert.match(skill, /inherit the[\s\S]*session model/i);
-});
-
 test("#given MultiAgentV2 transport #when members are added and bound #then task names and canonical agent paths form the address book", () => {
 	const tempRoot = createTeamRoot("omo-codex-teammode-v2-transport-");
 	try {
@@ -123,7 +83,6 @@ test("#given MultiAgentV2 transport #when members are added and bound #then task
 				{ id: "B", taskName: "mailbox_delivery", agentPath: "/root/mailbox_delivery", threadId: null, threadTitle: null, status: "active" },
 			],
 		);
-		assert.match(guide, /Transport:.*MultiAgentV2/i);
 		assert.match(guide, /members\[\]\.agentPath/);
 		assert.match(guide, /send_message/);
 		assert.match(guide, /followup_task/);
@@ -223,9 +182,7 @@ test("#given a V2 team #when the whole team is archived without a note #then no
 		const archiveEntry = team.log.find((entry) => entry.event === "archive");
 
 		assert.equal(team.status, "archived");
-		assert.doesNotMatch(result.stdout, /closed all members/i, "V2 has no runtime close/archive operation to claim");
 		assert.match(result.stdout, /no runtime archive operation/i);
-		assert.doesNotMatch(archiveEntry.detail, /all members closed/i);
 		assert.match(archiveEntry.detail, /no runtime archive operation/i);
 	} finally {
 		cleanupTeamRoot(tempRoot);
@@ -261,7 +218,6 @@ test("#given a codex_app team #when one member is archived #then the existing wo
 		const result = runTeam(tempRoot, "archive", "--team", "app-archive-member", "--id", "A");
 
 		assert.match(result.stdout, /archived member A/);
-		assert.doesNotMatch(result.stdout, /no runtime archive operation/i);
 	} finally {
 		cleanupTeamRoot(tempRoot);
 	}

package/packages/omo-codex/plugin/test/teammode-worktree.test.mjs

@@ -83,24 +83,21 @@ test("#given a worktree team #when worktree-add A #then a real git worktree on a
 		assert.match(result.stdout, /A/);
 		assert.match(result.stdout, /wait for the real Codex thread id/);
 		assert.match(result.stdout, /bind-thread before sending bootstrap/);
-		assert.doesNotMatch(result.stdout, /Tell that member to:/);
 	} finally {
 		cleanupTeamRoot(tempRoot);
 	}
 });
 
-test("#given worktree-add ran #when the field manual is regenerated #then it flips the member's guide to ISOLATION IS ON with the derived branch", () => {
+test("#given worktree-add ran #when the field manual is regenerated #then it shows the member's derived worktree branch", () => {
 	const tempRoot = createTeamRoot("omo-codex-teammode-wt-guide-");
 	try {
 		initGitRepo(tempRoot);
 		bootstrapTeam(tempRoot, "wt-guide", { worktree: false });
-		// before: no isolation, so the manual tells members to signal collisions instead
-		assert.match(readFileSync(join(teamDir(tempRoot, "wt-guide"), "guide.md"), "utf8"), /does not use isolated git worktrees/);
 
 		runTeam(tempRoot, "worktree-add", "--team", "wt-guide", "--id", "A");
 
+		// then - the regenerated guide carries the derived branch recorded in team state
 		const guide = readFileSync(join(teamDir(tempRoot, "wt-guide"), "guide.md"), "utf8");
-		assert.match(guide, /ISOLATION IS ON/);
 		assert.match(guide, /team\/wt-guide\/A/, "the member's row must show its derived worktree branch");
 	} finally {
 		cleanupTeamRoot(tempRoot);
@@ -225,7 +222,6 @@ test("#given a member branch that cannot be merged #when integrate runs #then it
 		const out = `${result.stdout}${result.stderr}`;
 		assert.notEqual(result.status, 0);
 		assert.match(out, /could not integrate|not something we can merge|ghost-branch-404/i, "must surface git's actual reason");
-		assert.doesNotMatch(out, /Conflicting files/, "a non-conflict failure must NOT be mislabeled as a merge conflict");
 	} finally {
 		cleanupTeamRoot(tempRoot);
 	}

package/packages/omo-codex/plugin/test/ultrawork-skill-pointer.test.mjs

@@ -33,7 +33,6 @@ test("#given synced skills #when the ultrawork skill is inspected #then it carri
 	const directive = readFileSync(join(root, "components", "ultrawork", "directive.md"), "utf8");
 	assert.match(skill, /^---\nname: ultrawork\n/);
 	assert.equal(skill.endsWith(directive), true);
-	assert.match(skill, /## 1\. Create the goal with binding success criteria/);
 });
 
 test("#given the built ultrawork CLI #when a ulw prompt is submitted #then a compact skill pointer is emitted", () => {
@@ -42,10 +41,8 @@ test("#given the built ultrawork CLI #when a ulw prompt is submitted #then a com
 	const output = JSON.parse(result.stdout);
 	const context = output.hookSpecificOutput.additionalContext;
 	assert.match(context, /^<ultrawork-mode>/);
-	assert.match(context, /First user-visible line this turn MUST be exactly:/);
 	assert.match(context, /create_goal/);
 	assert.equal(context.includes(skillPath), true);
-	assert.equal(context.includes("Tier triage"), false);
 	assert.equal(Buffer.byteLength(context, "utf8") < POINTER_MAX_BYTES, true);
 });
 

package/packages/omo-codex/plugin/test/ulw-plan-review-state-contract.test.mjs

@@ -54,15 +54,12 @@ function matchesCas(state, expected, clauses) {
 
 for (const surface of surfaces) {
 	test(`#given ${surface.name} #when deeper review becomes required before plan completion #then durable request state covers explicit and automatic review without inventing a digest`, async () => {
-		const skill = await readFile(surface.skillPath, "utf8");
 		const workflow = await readFile(surface.workflowPath, "utf8");
 		const contract = readJsonContract(workflow, "ulw-plan-review-request-state-contract");
 
 		assert.equal(contract.transition, "replace");
 		assert.equal(contract.phase, "review_requested");
 		assert.deepEqual(contract.applies_when, ["explicit_review_modifier_before_complete_plan", "intent=unclear_and_nontrivial"]);
-		assert.match(skill, /Include `--review-required`[\s\S]{0,180}non-Trivial UNCLEAR/);
-		assert.match(workflow, /--draft-only/);
 		assert.equal(contract.atomic, true);
 		assert.equal(contract.review_required, true);
 		assert.equal(contract.plan_path, ".omo/plans/<slug>.md");