Enhance auto briefing functionality and documentation for v2 by Looted · Pull Request #158 · Looted/kibi

Looted · 2026-04-23T14:57:13Z

No description provided.

…riefing-v2

…seed-ID sourcing

…nsform hook - Integrate computeBriefIntent into file.edited handler to assess eligible risk classes - Store BriefingRuntimeResult in autoBriefResults map keyed by fingerprint - Send toast via sendToast when fetchBriefingResult resolves - Pass autoBriefResult into buildPrompt context for prompt block rendering - Add PromptContext.autoBriefResult field - Add computeBriefIntent traceability link to REQ-opencode-kibi-briefing-v2 - Add 496 lines of integration tests covering all eligibility/ineligibility paths

…rror handling

…mbol docs Replace schema-invalid supersedes edges (scenario→scenario, test→test) with relates_to in SCEN-opencode-kibi-briefing-v2 and TEST-opencode-kibi-briefing-v2. Remove plain string links arrays from SYM-cli-status-pre-first-sync-test and SYM-mcp-cli-help-test that were being mis-imported as invalid implements edges; executable_for relationships are preserved.

Copilot

Pull request overview

This PR introduces v2 “auto-briefing” for the OpenCode plugin: it fetches citation-backed briefing guidance asynchronously on file.edited for authoritative risky edits, renders a dedicated “Kibi briefing available” guidance block in the prompt, and updates supporting docs/tests across opencode and mcp.

Changes:

Add auto-brief intent derivation + background worker runtime (brief-intent, briefing-runtime) and integrate it into the OpenCode plugin event/transform flow.
Update prompt rendering to support auto-brief blocks, cue suppression, and bullet-budget behavior under completion reminders.
Expand/adjust tests and documentation for v2 briefing behavior; add toast utility and update MCP tool tests.

Reviewed changes

Copilot reviewed 26 out of 26 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
packages/opencode/src/index.ts	Triggers auto-brief fetch on `file.edited`, stores results/toast dedupe, passes cached result into prompt transform.
packages/opencode/src/prompt.ts	Renders auto-brief blocks, adjusts bullet budgeting when reminders append, suppresses cue/micro-brief as appropriate.
packages/opencode/src/brief-intent.ts	Computes eligibility + fingerprint + seed IDs for auto-briefing.
packages/opencode/src/briefing-runtime.ts	Implements worker session prompting, payload parsing, caching, and runtime result normalization.
packages/opencode/src/toast.ts	Centralizes toast capability detection + sending (showToast vs legacy toast).
packages/opencode/src/startup-notifier.ts	Refactors startup toast handling to use the new toast utility.
packages/opencode/tests/prompt.test.ts	Adds prompt-level tests for auto-brief rendering/cue suppression/budgeting.
packages/opencode/tests/index.test.ts	Adds integration tests covering event-triggered fetch, toast behavior, dedupe, and transform rendering.
packages/opencode/tests/hook-contract.test.ts	Tightens hook output policy to avoid embedding briefing execution/payload artifacts.
packages/opencode/tests/brief-intent.test.ts	Adds unit tests for intent eligibility/fingerprint/seed derivation behavior.
packages/opencode/tests/briefing-auto-render.test.ts	Adds runtime-level tests for worker prompting, parsing, caching, and concurrency dedupe.
packages/opencode/tests/agent-surface-policy.test.ts	Registers v2 requirement/scenario/test docs in the surface policy allowlist.
packages/mcp/tests/tools/briefing-generate.test.ts	Ensures “no briefing” responses include a consistent text content payload.
documentation/**	Adds v2 REQ/SCEN/TEST docs; deprecates v1 docs.
packages/opencode/package.json, packages/mcp/package.json, CHANGELOG.md	Version/changelog updates aligned with the new feature set.

Comments suppressed due to low confidence (1)

packages/opencode/src/index.ts:378

The cache logger reports a miss even when cache.isSatisfied(cacheKey) is true for auto-brief risk classes, because execution falls through to the unconditional "cache_state: "miss"" log after the hit branch. This makes telemetry misleading and can also cause cache-hit behavior to execute twice. Consider restructuring this block so the miss log only runs when the cache is not satisfied (e.g., else), and if auto-brief work must still run on cache hits, run that work in the hit branch and return afterward.

    // Cache check: after critical signals have been emitted
    if (cache.isSatisfied(cacheKey)) {
      logger.info("smart-enforcement.cache", {
        event: "smart_enforcement_cache",
        cache_hit: true,
        cache_state: "hit",
        file: filePath,
        risk_class: effectiveRiskClass,
        posture: posture.state,
        posture_state: posture.state,
      });
      if (!isAutoBriefRisk) {
        return;
      }
    }

    logger.info("smart-enforcement.cache", {
      event: "smart_enforcement_cache",
      cache_hit: false,
      cache_state: "miss",
      file: filePath,
      risk_class: effectiveRiskClass,
      posture: posture.state,
      posture_state: posture.state,
    });

+  if (params.seedIds !== undefined) {
+    return params.seedIds.slice(0, 3);
+  }
+
+  return getSourceLinkedRequirementIds(
+    params.workspaceRoot,
+    params.editedFilePath,
+  ).slice(0, 3);


+        const autoBriefResult = (() => {
+          if (lastRiskClass == null || lastEditedFilePath == null) {
+            return undefined;
+          }
+
+          const intentResult = computeBriefIntent({
+            riskClass: lastRiskClass,
+            posture: posture.state,
+            maintenanceDegraded,
+            editedFile: lastEditedFilePath,
+            worktreeRoot: input.worktree,
+            branch: currentBranch,
+          });
+
+          return autoBriefResults.get(intentResult.fingerprint);
+        })();


+19#YT|> **Note**: This test doc is DEPRECATED and superseded by TEST-opencode-kibi-briefing-v2. 
+20#YT|> Historical verification for v1 cue-driven briefings remains documented below.


+20#YT|> **Note**: This scenario is DEPRECATED and superseded by SCEN-opencode-kibi-briefing-v2. 
+21#YT|> It documents the historical v1 cue-driven behavior.


+31#YT|> **Note**: This requirement is DEPRECATED and superseded by REQ-opencode-kibi-briefing-v2. 
+32#YT|> It remains here for historical context and to document the v1 cue-driven contract.


Looted · 2026-04-23T15:44:04Z

@copilot apply changes based on the comments in this thread in a separate PR

Copilot · 2026-04-23T15:44:14Z

@Looted I've opened a new pull request, #159, to work on those changes. Once the pull request is ready, I'll request review from you.

- brief-intent.ts: normalize editedFilePath to absolute path before passing to getSourceLinkedRequirementIds() to fix silent [] return when file.edited emits a relative path - index.ts: cache lastBriefFingerprint from file.edited handler; system.transform now looks up the cached fingerprint instead of recomputing computeBriefIntent() (avoids repeated symbols.yaml I/O) - Remove accidental #YT|> annotation prefixes from REQ, SCEN, and TEST v1 documentation markdown files Agent-Logs-Url: https://github.com/Looted/kibi/sessions/48b5f91e-68ec-4a84-98ad-f37766084caf Co-authored-by: Looted <6255880+Looted@users.noreply.github.com>

[WIP] Enhance auto briefing functionality and documentation for v2

Looted · 2026-04-23T15:55:55Z

@copilot I might have merged before you addressed all the comments. Could you open another PR with the rest?

Copilot · 2026-04-23T15:56:04Z

@Looted I've opened a new pull request, #160, to work on those changes. Once the pull request is ready, I'll request review from you.

Looted added 16 commits April 23, 2026 10:38

docs(opencode): evolve kibi briefing contract to v2 auto-show

a52f94b

fix(docs): remove corrupted line-id markers from SCEN-opencode-kibi-b…

d208668

…riefing-v2

test(opencode): add briefing intent helper matrix

bfd805b

refactor(opencode): improve brief-intent helper with ReadonlySet and …

c097db6

…seed-ID sourcing

refactor(opencode): centralize toast capability handling

384557b

feat(opencode): add auto-brief runtime helper

9862a8e

feat(opencode): trigger auto briefing from event path

40a19ef

feat(opencode): surface auto briefing in prompt guidance

916d2eb

fix(opencode): deduplicate toast emission per fingerprint

3069d59

test(opencode): cover auto briefing runtime paths

f538e6f

docs(opencode): describe auto briefing v2

f9258c6

fix(opencode): remove dead code, fix budget overflow, add sendToast e…

4b87b7d

…rror handling

Version bump, added opencode briefing automation

a53c9ac

Merge branch 'develop' of https://github.com/Looted/kibi into develop

a6c6d58

Copilot AI review requested due to automatic review settings April 23, 2026 14:57

Copilot started reviewing on behalf of Looted April 23, 2026 15:04 View session

Copilot AI reviewed Apr 23, 2026

View reviewed changes

Initial plan

d3c2e43

Copilot AI mentioned this pull request Apr 23, 2026

Fix auto-briefing path normalization, fingerprint caching, and doc annotation artifacts #159

Merged

Copilot AI and others added 2 commits April 23, 2026 15:48

Merge pull request #159 from Looted/copilot/sub-pr-158

120debb

[WIP] Enhance auto briefing functionality and documentation for v2

Copilot AI mentioned this pull request Apr 23, 2026

Address remaining PR #158 review comments (all fixed in de6db84) #160

Closed

Looted merged commit 18b491c into master Apr 23, 2026
6 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Enhance auto briefing functionality and documentation for v2#158

Enhance auto briefing functionality and documentation for v2#158
Looted merged 19 commits intomasterfrom
develop

Looted commented Apr 23, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Looted commented Apr 23, 2026

Uh oh!

Copilot AI commented Apr 23, 2026

Uh oh!

Looted commented Apr 23, 2026

Uh oh!

Copilot AI commented Apr 23, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

		19#YT\|> Note: This test doc is DEPRECATED and superseded by TEST-opencode-kibi-briefing-v2.
		20#YT\|> Historical verification for v1 cue-driven briefings remains documented below.

		20#YT\|> Note: This scenario is DEPRECATED and superseded by SCEN-opencode-kibi-briefing-v2.
		21#YT\|> It documents the historical v1 cue-driven behavior.

		31#YT\|> Note: This requirement is DEPRECATED and superseded by REQ-opencode-kibi-briefing-v2.
		32#YT\|> It remains here for historical context and to document the v1 cue-driven contract.

Conversation

Looted commented Apr 23, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Looted commented Apr 23, 2026

Uh oh!

Copilot AI commented Apr 23, 2026

Uh oh!

Looted commented Apr 23, 2026

Uh oh!

Copilot AI commented Apr 23, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants