@sireai/optimus 0.1.9 → 0.1.10

package/task-harnesses/bugfix/STANDARD.md

@@ -19,30 +19,45 @@
 ## Validation policy
 
 - A claimed fix requires validation.
-- Prefer stronger evidence before weaker evidence.
-- Report the strongest level reached and why stronger levels were unavailable.
+- Prefer the highest-reliability validation that is feasible now, not the cheapest one.
+- If a stronger level is blocked, state the blocker and downgrade explicitly.
+- Report exactly one strongest token and keep the method/result details below it.
+- Also report exactly one validation grade `V1` to `V5`.
 
-1. `L4 functional`: real device, simulator, or another directly runnable environment
-2. `L3 self-check`: local unit tests, targeted custom tests, lightweight scenario injection
-3. `L2 build`: relevant compile target, module build, or targeted test task
-4. `L1 code evidence`: static reasoning, call-chain review, diff review
+Reliability and cost order:
 
-### Android order
+1. `V5`: `device_verified` - highest confidence, highest cost
+2. `V4`: `simulator_verified`, `scenario_verified` - very high or high confidence, high or medium-high cost
+3. `V3`: `regression_tests_passed`, `unit_tests_passed` - medium-high or medium confidence, medium or medium-high cost
+4. `V2`: `module_build_passed`, `compile_passed`, `targeted_tests_passed` - partial executable proof, medium to lower confidence, low-medium to medium cost
+5. `V1`: `code_reviewed` - lowest confidence, lowest cost
 
-1. Prefer real-device or simulator validation when `adb` and runnable targets exist.
-2. Otherwise prefer local tests or scenario injection.
-3. Use device-side automated checks only when local validation cannot cover the behavior.
-4. Fall back to compile validation such as `compileDebugKotlin` or relevant unit tests.
-5. Use code-only validation only when all stronger forms are unavailable.
+Token contract:
+
+- `device_verified` -> `V5`: real device, real business path, fix behavior observed
+- `simulator_verified` -> `V4`: simulator/emulator, real business path, fix behavior observed
+- `scenario_verified` -> `V4`: directly runnable real path or end-to-end scenario executed in a near-real environment
+- `regression_tests_passed` -> `V3`: multiple relevant existing regression/integration cases passed
+- `unit_tests_passed` -> `V3`: real project unit tests passed
+- `module_build_passed` -> `V2`: real module/package build passed
+- `compile_passed` -> `V2`: real compile target passed
+- `targeted_tests_passed` -> `V2`: targeted harness, stub, mock, injected script, minimal API-surface check, or one-off focused executable proof passed
+- `code_reviewed` -> `V1`: no executable validation succeeded; conclusion relies on code/log/evidence review only
+
+Never overstate:
+
+- Stub, mock, temporary harness, minimal API surface, injected script, or ad hoc executable proof must be `targeted_tests_passed`, not `scenario_verified`.
+- Real compile/build did not pass: do not claim `compile_passed` or `module_build_passed`.
+- No real device/simulator path was exercised: do not claim `device_verified`, `simulator_verified`, or `scenario_verified`.
 
 ## Closure policy
 
 - Close as fix only when analysis, code changes, validation evidence, and residual-risk understanding are credible.
 - Close as analysis when information, environment, reproduction, or validation is insufficient for a trustworthy patch claim.
-- If code changed but validation reached only `L2` or `L1`, describe it as a repair candidate, not a verified fix.
-- If the issue is interaction, crash, device, integration, or resource related and validation stayed at `L2`, state what stronger environment or tooling was missing.
+- If code changed but fix validation stayed at `V2` or `V1`, describe it as a repair candidate, not a verified fix.
+- If the issue is interaction, crash, device, integration, or resource related and fix validation stayed at `V2`, state what stronger environment or tooling was missing.
 - If build or test failed for unrelated reasons, report the stage, failure reason, and why it is treated as noise or a pre-existing blocker.
-- If only `L1` evidence exists, do not submit a formal patch claim; close as analysis.
+- If only `V1` evidence exists, do not submit a formal verified-fix claim; close as analysis unless a repair candidate is still justified.
 - Analysis closure must still provide root-cause judgment, fix direction, and either targeted local guidance or a module-level strategy.
 
 ## Runtime contract
@@ -81,7 +96,8 @@
 - Before writing `result.md`, determine `Closure Level`, then follow exactly one language mode:
   - `Verified Fix` or `Repair Candidate`: Patch Closure Mode; all narrative sections are English
   - `Analysis Only`: Analysis Closure Mode; narrative sections are Chinese
-- `Validation Summary` stays a single English token in all cases.
+- `Reproduction` / `复现情况` uses the compact form `<token> (R*)` when grade is known.
+- `Fix Validation` / `修复验证` uses the compact form `<token> (V*)` when grade is known.
 - Use repository-relative code paths only; never use absolute local paths.
 - Commands, logs, stack traces, API errors, and identifiers may stay in their original language when needed.
 - If closure is patch-related and any narrative field is Chinese, the output is invalid and must be rewritten.
@@ -91,14 +107,25 @@
 Do not rename these downstream-consumed keys:
 
 - English patch-mode keys:
-  - `Validation Summary`
+  - `Root Cause`
+  - `Fix`
+  - `Reproduction`
+  - `Fix Validation`
+  - `Impact Check`
+  - `Confidence`
+  - `Blocking Point`
   - `Strongest Current Conclusion`
-  - `Analysis Summary`
   - `Key Evidence`
   - `Recommended Action`
   - `Analysis Doc URL`
 - Chinese analysis-mode keys:
-  - `验证摘要`
+  - `根因摘要`
+  - `修复建议`
+  - `复现情况`
+  - `修复验证`
+  - `影响评估`
+  - `确定性`
+  - `阻塞点`
   - `当前最强结论`
   - `分析摘要`
   - `关键证据`
@@ -111,10 +138,25 @@ Keep exact capitalization and wording.
 
 At minimum, `result.md` must include:
 
+- one compact summary block for downstream delivery consumers
+  Patch mode fields: `Root Cause`, `Fix`, `Reproduction`, `Fix Validation`, `Impact Check`, `Confidence`, `Blocking Point`
+  Analysis mode fields: `根因摘要`, `修复建议`, `复现情况`, `修复验证`, `影响评估`, `确定性`, `阻塞点`
+  Keep each summary value dense and short enough for comment/card reuse.
+- a reproduction summary as one high-density English token
+  Allowed values: `naturally_reproduced`, `induced_reproduced`, `historical_evidence_matched`, `not_reproduced`
+- a reproduction grade inside `Reproduction` / `复现情况`
+  Allowed values: `R1`, `R2`, `R3`, `R4`
 - a validation summary as one high-density English token
   Allowed values: `device_verified`, `simulator_verified`, `scenario_verified`, `unit_tests_passed`, `targeted_tests_passed`, `regression_tests_passed`, `compile_passed`, `module_build_passed`, `code_reviewed`
   Forbidden generic values: `validation_completed`, `tests_passed`, `verified`, `passed`, `done`
-  If multiple validations were performed, report only the strongest one.
+  If multiple validations were performed, report only the strongest one by the reliability order above.
+- a validation grade inside `Fix Validation` / `修复验证`
+  Allowed values: `V1`, `V2`, `V3`, `V4`, `V5`
+  It must match the selected validation token.
+- an impact-check summary token
+  Allowed values: `neighbor_paths_checked`, `partial_neighbor_check`, `not_checked`
+- a confidence token
+  Allowed values: `C1`, `C2`, `C3`, `C4`
 - problem summary and impact scope
 - category: functional, stability, performance, or compatibility
 - reproduction likelihood: always, high, low, or unknown
@@ -145,30 +187,42 @@ At minimum, `result.md` must include:
 ```md
 # Bugfix Result
 
-## Summary
+## Delivery Summary
+- Root Cause:
+- Fix:
+- Reproduction:
+- Fix Validation:
+- Impact Check:
+- Confidence:
+- Blocking Point: `None` if not blocked
+
+## Detail
+
+### Summary
 - Problem:
 - Impact:
 - Category: Functional / Stability / Performance / Compatibility
 - Reproduction Likelihood: Always / High / Low / Unknown
 
-## Root Cause
+### Root Cause
 - Strongest Current Conclusion:
 - Key Evidence:
 - Relevant Code Locations:
 
-## Change
+### Change
 - Closure Level: Verified Fix / Repair Candidate
 - Patch Notes:
 - Fix Strategy:
 - Blocking Point: `None` if not blocked
 
-## Validation
+### Validation
 - Validation Summary: exactly one short English token, strongest validation only
+- Validation Grade: exactly one token, `V1` to `V5`, matching `Validation Summary`
 - Method:
 - Result:
 - Unverified Items:
 
-## Risks
+### Risks
 - Residual Risk:
 - Recommended Action:
 ```
@@ -190,29 +244,41 @@ Use the following Chinese output structure exactly:
 ```md
 # 缺陷分析结果
 
-## 问题概述
+## 交付摘要
+- 根因摘要:
+- 修复建议:
+- 复现情况: `<token> (R*)` when grade is known
+- 修复验证: `<token> (V*)` when grade is known
+- 影响评估:
+- 确定性:
+- 阻塞点:
+
+## 详细分析
+
+### 问题概述
 - 问题:
 - 影响:
 - 分类: 功能 / 稳定性 / 性能 / 兼容性
 - 复现概率: 必现 / 高概率 / 低概率 / 未知
 
-## 分析结论
+### 分析结论
 - 当前最强结论:
 - 分析摘要:
 - 关键证据:
 - 相关代码位置:
 
-## 修复判断
+### 修复判断
 - Closure Level: Analysis Only
 - 阻塞点:
 - 建议动作:
 
-## 验证情况
+### 验证情况
 - 验证摘要: exactly one short English token, strongest validation only
+- 验证等级: exactly one token, `V1` to `V5`, matching `验证摘要`
 - 已验证内容:
 - 未验证内容:
 
-## 风险
+### 风险
 - 主要风险:
 - 建议动作:
 - 分析文档链接:
@@ -236,6 +302,8 @@ Patch closure examples:
 - If code changed, ensure `result.md` and `patch.diff` do not contradict each other.
 - If important code changed, ensure explanatory comments are present.
 - If validation was performed, ensure claims are not overstated.
+- Ensure validation token matches the strongest proof actually executed, not the intended proof.
+- Ensure `Delivery Summary` / `交付摘要` is present and consistent with the detailed sections below it.
 - Distinguish confirmed facts from inference.
 - For patch closure, ensure `Strongest Current Conclusion`, `Key Evidence`, and `Recommended Action` are English before returning.
 
@@ -248,4 +316,9 @@ Patch closure examples:
 - required downstream field names are exact
 - patch closure does not emit `Analysis Summary`
 - analysis closure does not mix in patch-delivery sections
-- `Validation Summary` or `验证摘要` is exactly one strongest English token
+- `Reproduction` or `复现情况` uses exactly one allowed token and an `R*` grade when grade is present
+- `Fix Validation` or `修复验证` uses exactly one strongest allowed validation token and a matching `V*` grade when grade is present
+- `Impact Check` or `影响评估` uses exactly one allowed token
+- `Confidence` or `确定性` uses exactly one allowed token
+- detailed validation section still contains `Validation Summary` / `验证摘要`
+- detailed validation section still contains `Validation Grade` / `验证等级`

package/embedded-skills/shared/video-keyframe-analyzer/references/encountered-problems.md

@@ -1,12 +0,0 @@
-# Encountered Problems
-
-Known operational issues from real Jira recording analysis:
-
-- Jira attachment links may require browser/CAS auth; direct curl can return a login page.
-- Browser downloads may appear first as hidden temporary files before the final filename is visible.
-- Interactive shell environment variables may not reach non-interactive task commands.
-- Cloud video understanding can be blocked by network sandboxing.
-- Remote video processing can be slow even for short recordings.
-- Remote polling can fail with transient partial reads.
-
-The local keyframe workflow avoids most of these by reducing video analysis to local image artifacts.

package/embedded-skills/shared/video-keyframe-analyzer/references/triage-checklist.md

@@ -1,48 +0,0 @@
-# Triage Checklist
-
-Use this compact structure after reviewing keyframes.
-
-## Summary
-
-- User flow
-- Visible failure
-- Post-failure state
-
-## Timeline
-
-- `00:00-00:03`: entry state
-- `00:03-00:08`: visible user action or transition
-- `00:08-00:10`: anomaly
-- `00:10+`: post-failure behavior
-
-## Repro Steps
-
-Infer only what the recording supports:
-
-1. Open app and enter the visible module.
-2. Perform the visible interaction.
-3. Observe the incorrect outcome.
-
-## Observed vs Expected
-
-- Observed: what the UI actually did.
-- Expected: what should happen in a healthy flow.
-
-## Evidence
-
-Useful visual evidence includes:
-
-- app returns to launcher unexpectedly
-- relaunch lands on login or wrong page
-- dialog remains visible after state changes
-- loading spinner or white screen stays unchanged
-- expected toast, navigation, or dismissal never appears
-
-## Open Questions
-
-Keep speculation separate:
-
-- account or environment state
-- network condition
-- exact trigger if not visible
-- whether the issue reproduces across devices or accounts