feat(agents): add Task Reviewer and expand RPI to 4-phase workflow#277
feat(agents): add Task Reviewer and expand RPI to 4-phase workflow#277agreaves-ms merged 45 commits intomainfrom
Conversation
Dependency Review✅ No vulnerabilities or license issues or OpenSSF Scorecard issues found.Scanned FilesNone |
There was a problem hiding this comment.
Pull request overview
This PR introduces a significant expansion of the RPI (Research-Plan-Implement) workflow by adding a fourth Review phase, transforming it into Research-Plan-Implement-Review. The review phase enables validation against documented specifications, convention compliance checking, and structured iteration loops back to earlier phases when needed.
Changes:
- Adds Task Reviewer agent for implementation validation with findings categorized by severity (Critical, Major, Minor)
- Adds memory agent for conversation context persistence across session resets with save/continue/incremental modes
- Adds doc-ops agent for autonomous documentation quality assurance separate from RPI workflow
- Transforms rpi-agent into fully autonomous 5-phase orchestrator (Research → Plan → Implement → Review → Discover) dispatching task-* agents via runSubagent
- Adds handoff frontmatter to all RPI task agents enabling seamless phase transitions via UI buttons
- Removes task-implementation.instructions.md (consolidated into task-implementor agent)
- Updates all RPI documentation from 3-phase to 4-phase workflow
- Standardizes date formatting to
{{YYYY-MM-DD}}syntax across agents and prompts - Removes agent references from ADO prompts for flexibility
- Updates task-planner to create 2 files instead of 3 (removed generated implementation prompt)
Reviewed changes
Copilot reviewed 41 out of 41 changed files in this pull request and generated no comments.
Show a summary per file
| File | Description |
|---|---|
| extension/package.json | Registers new agents (doc-ops, memory, task-reviewer), prompts (checkpoint, doc-ops-update, rpi, task-review), removes task-implementation instructions, updates descriptions |
| docs/rpi/why-rpi.md | Updates to describe 4-phase workflow with Task Reviewer, adds review phase benefits and iteration patterns |
| docs/rpi/using-together.md | Adds Phase 4: Review section with validation workflow, iteration loops, and handoff button documentation |
| docs/rpi/task-reviewer.md | New comprehensive guide for Task Reviewer usage, severity levels, and iteration paths |
| docs/rpi/task-researcher.md | Adds handoff button tip, updates date format placeholders, minor formatting improvements |
| docs/rpi/task-planner.md | Updates for 2-file output (removed prompt generation), adds handoff button tip |
| docs/rpi/task-implementor.md | Updates handoff to review phase, removes prompt deletion requirement |
| docs/rpi/README.md | Updates overview for 4-phase workflow, adds Task Reviewer phase description |
| docs/contributing/prompts.md | Updates date format placeholders to {{YYYY-MM-DD}} |
| docs/contributing/custom-agents.md | Adds handoffs frontmatter documentation with examples and field specifications |
| .github/prompts/task-review.prompt.md | New prompt for initiating review with scope parameters and artifact discovery |
| .github/prompts/rpi.prompt.md | New prompt for autonomous RPI workflow with autonomy modes (auto/partial/false) |
| .github/prompts/checkpoint.prompt.md | New prompt for memory file operations (save/continue/incremental modes) |
| .github/prompts/doc-ops-update.prompt.md | New prompt for documentation quality assurance with scope filtering |
| .github/prompts/task-research.prompt.md | Adds important requirements section, updates date format placeholders |
| .github/prompts/task-plan.prompt.md | Adds important requirements, updates for 2-file creation |
| .github/prompts/task-implement.prompt.md | Adds important requirements section, updates date format placeholders |
| .github/prompts/prompt-refactor.prompt.md | Adds important requirements section for thoroughness |
| .github/prompts/prompt-build.prompt.md | Adds important requirements section, refactoring guidance |
| .github/prompts/pull-request.prompt.md | Removes agent reference for flexibility |
| .github/prompts/ado-*.prompt.md | Removes agent references from all ADO prompts |
| .github/agents/task-reviewer.agent.md | New agent with 5-phase protocol (Discovery, Extraction, Validation, Follow-Up, Completion) and subagent dispatch |
| .github/agents/memory.agent.md | New agent with 3-phase protocol (Detect, Save, Continue) for conversation persistence |
| .github/agents/doc-ops.agent.md | New autonomous agent with 5-phase protocol for documentation QA |
| .github/agents/task-researcher.agent.md | Adds handoff to task-planner, updates response format with emoji header, structured completion handoff |
| .github/agents/task-planner.agent.md | Adds handoff to task-implementor, updates for 2-file output, structured completion handoff |
| .github/agents/task-implementor.agent.md | Adds handoff to task-reviewer, updates response format, adds deviations tracking |
| .github/agents/rpi-agent.agent.md | Major rewrite as autonomous 5-phase orchestrator with autonomy modes and discovery phase |
| .github/agents/prompt-builder.agent.md | Adds handoff buttons for workflow transitions, updates iteration logic |
| .github/agents/hve-core-installer.agent.md | Updates VS Code settings to include chat.agentFilesLocations, removes .copilot-tracking/prompts |
| .github/agents/gen-data-spec.agent.md | Updates date format placeholders to {{YYYY-MM-DD}} |
| .github/instructions/task-implementation.instructions.md | Removed (consolidated into task-implementor agent) |
| .github/instructions/prompt-builder.instructions.md | Adds user-facing response guidance for markdown link formatting |
| .github/instructions/ado-get-build-info.instructions.md | Updates date format placeholders to {{YYYY-MM-DD}} |
| .github/instructions/README.md | Removes task-implementation.instructions.md references |
| .github/copilot-instructions.md | Removes .copilot-tracking/prompts/ directory reference |
| .github/CUSTOM-AGENTS.md | Updates date formats, prompt counts (2 files not 3 for planner) |
WilliamBerryiii
left a comment
There was a problem hiding this comment.
Approved with suggestions. LFGTM - thanks for knocking this out and can't wait to run with the new reviewer.
…pdate related documentation - consolidate task implementation guidance into existing agents - enhance clarity in task planning prompts - update package.json to reflect instruction removal 🔧 - Generated by Copilot
🔍 - Generated by Copilot
… reviews - specify thoroughness and precision in validation - allow subagents to ask clarifying questions - emphasize continuous log updates during validation - repeat steps for additional context as needed 🔍 - Generated by Copilot
- add clarity on thoroughness and accuracy across prompts - emphasize avoiding verbosity and ensuring completeness in documentation - encourage iterative refinement of instructions and examples 🔒 - Generated by Copilot
…tion reviews - update handoff instructions for review findings - clarify summary and next steps for user handoff - improve consistency in important requirements across prompts 🔒 - Generated by Copilot
Add VS Code Custom Agent handoffs: declarations enabling seamless transitions between RPI workflow phases via handoff buttons. Agent updates: - task-researcher: handoff to task-planner - task-planner: handoff to task-implementor - task-implementor: handoff to task-reviewer - task-reviewer: handoffs to task-researcher and task-planner Documentation updates: - custom-agents.md: document handoffs field in Optional Fields section - using-together.md: update workflow diagram, add Handoff Buttons section
- introduce new rpi.prompt.md for RPI agent workflow - update task-planner and task-implement prompts for clarity - refine task-review prompt to enhance validation process 🔄 - Generated by Copilot
- update research, planning, implementation, and review phases to include discovered instructions and skills - improve clarity on subagent responsibilities and artifact references 🔍 - Generated by Copilot
🔒 - Generated by Copilot
… and response format - present findings summary with severity counts in a table - include review log file path for detailed reference - provide numbered handoff steps based on review outcome - add user interaction patterns for rework and follow-up items 🔍 - Generated by Copilot
… RPI agent - streamline handoff options for user interactions - clarify response standards for completion summaries 🔧 - Generated by Copilot
- implement memory file structure for session continuity - define phases for detection, saving, and continuing memory - outline user interaction and response formats for memory operations 🔒 - Generated by Copilot
…n behavior - add continuation behavior for assessing conversation context - define response standards for user-facing communication - implement phase transition updates and completion patterns 🔄 - Generated by Copilot
…for memory context - add incremental mode for mid-session progress saves - update operation steps to include task overview and current state - clarify inputs and execution for improved user interaction 🔒 - Generated by Copilot
…mpt builder 🔧 - Generated by Copilot
… sessions - add identification of custom agents invoked - include notes on custom agents for workflow continuity 🔄 - Generated by Copilot
… management 🔄 - Generated by Copilot
…nce continuation options 🔍 - Generated by Copilot
…vels - add auto parameter for controlling autonomy - update continue and suggest parameters for clarity - refine input handling descriptions for better user guidance 🔧 - Generated by Copilot
…otocol - introduce handoff for continuing with RPI - provide instructions for switching agents in VS Code - enhance user guidance for session continuity 🚀 - Generated by Copilot
- update task researcher, planner, implementor, and reviewer guides - clarify the new review phase and its importance in the workflow - adjust artifact summaries and common patterns for clarity - revise commit messages and instructions for better user guidance 📚 - Generated by Copilot
… definition - refine description for documentation quality assurance - adjust argument-hint to remove unnecessary options - enhance input descriptions for better understanding - clarify scope definition for documentation processing 🔧 - Generated by Copilot
…ion quality assurance 🔒 - Generated by Copilot
…d roadmap documents refactor(docs): enhance readability of task reviewer and using together documentation refactor(docs): standardize table layouts in why-rpi documentation 🔧 - Generated by Copilot
… and human reviewers 🔒 - Generated by Copilot
e3c6bf3 to
3003256
Compare
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #277 +/- ##
==========================================
+ Coverage 28.80% 28.87% +0.07%
==========================================
Files 14 14
Lines 2736 2736
==========================================
+ Hits 788 790 +2
+ Misses 1948 1946 -2
Flags with carried forward coverage won't be shown. Click here to find out more. 🚀 New features to boost your workflow:
|
🤖 I have created a release *beep* *boop* --- ## [2.0.0](hve-core-v1.1.0...hve-core-v2.0.0) (2026-01-28) ### ⚠ BREAKING CHANGES * **agents:** add Task Reviewer and expand RPI to 4-phase workflow ([#277](#277)) ### ✨ Features * **agents:** add hve-core-installer agent to extension package ([#297](#297)) ([c0e48c6](c0e48c6)) * **agents:** add Task Reviewer and expand RPI to 4-phase workflow ([#277](#277)) ([ae76cab](ae76cab)) * **build:** add code coverage reporting to Pester workflow ([#230](#230)) ([a34822a](a34822a)) * **docs:** add GOVERNANCE.md for OSSF Silver Badge compliance ([#235](#235)) ([b0e752c](b0e752c)) * **docs:** add ROADMAP.md for OSSF Silver badge compliance ([#238](#238)) ([4a41c16](4a41c16)) * **mcp:** add MCP server configuration guidance and installer enhancements ([#225](#225)) ([0bce418](0bce418)) * **scripts:** add YAML linting with actionlint ([#234](#234)) ([d9301f9](d9301f9)) * **security:** add OpenSSF Scorecard workflow and badge ([#271](#271)) ([7c6d788](7c6d788)) * **skills:** add video-to-gif conversion skill with FFmpeg two-pass optimization ([#247](#247)) ([8d65c42](8d65c42)) * **tests:** add Pester tests for LintingHelpers and Validate-MarkdownFrontmatter ([#197](#197), [#198](#198)) ([#205](#205)) ([51ae563](51ae563)) ### 🐛 Bug Fixes * **build:** detect table formatting changes via git diff ([#261](#261)) ([985eee0](985eee0)) * **build:** disable MD024 lint rule in CHANGELOG for release-please ([#220](#220)) ([971df94](971df94)) * **build:** quote shell variables and group redirects in workflow files ([#299](#299)) ([3372509](3372509)) * **build:** resolve scorecard badge and workflow security issues ([#301](#301)) ([aeaed13](aeaed13)) * **extension:** remove frontmatter from README and exclude from markdown linting ([#223](#223)) ([4272529](4272529)) * **instructions:** quote applyTo glob pattern for YAML compatibility ([#216](#216)) ([085199c](085199c)) * **scripts:** add FooterExcludePaths parameter to frontmatter validation ([#334](#334)) ([64db98d](64db98d)) * **scripts:** add GHSA word and logs/ exclusion to cspell config ([#214](#214)) ([5c99b3f](5c99b3f)) * **scripts:** correct type assertions in Invoke-YamlLint.Tests.ps1 ([#332](#332)) ([af7050d](af7050d)) * **scripts:** eliminate false positives in dependency pinning npm pattern ([#273](#273)) ([ccbdfa3](ccbdfa3)) * **security:** add artifact attestation for signed releases ([#257](#257)) ([c52d6e2](c52d6e2)) * standardize markdown footers and complete frontmatter ([#217](#217)) ([b4e7556](b4e7556)) ### 📚 Documentation * add OpenSSF Best Practices Passing badge to README ([#239](#239)) ([91bc529](91bc529)) * **architecture:** add architecture documentation and value proposition ([#252](#252)) ([0e4b02f](0e4b02f)) * **contributing:** add testing requirements for OSSF compliance ([#254](#254)) ([4db1a18](4db1a18)) * **docs:** add enterprise status badges to README header ([#270](#270)) ([ccb68a4](ccb68a4)) * **security:** add security assurance case and threat model for OSSF Silver ([#259](#259)) ([a390e26](a390e26)) ### ♻️ Refactoring * **application:** wrap execution with try blocks, ensure proper … ([#296](#296)) ([35c4417](35c4417)) * **scripts:** extract frontmatter validation to testable module ([#293](#293)) ([4e8707e](4e8707e)) * **scripts:** extract pure functions for Pester testability ([#221](#221)) ([d40e742](d40e742)) ### 🔧 Maintenance * **deps-dev:** bump cspell from 9.4.0 to 9.6.0 in the npm-dependencies group ([#208](#208)) ([855914b](855914b)) * **deps-dev:** bump cspell from 9.6.0 to 9.6.1 in the npm-dependencies group ([#294](#294)) ([1e45ad6](1e45ad6)) * **deps:** bump actions/setup-node from 6.1.0 to 6.2.0 in the github-actions group ([#209](#209)) ([c4c69e2](c4c69e2)) * **deps:** bump the github-actions group with 4 updates ([#295](#295)) ([d8337b8](d8337b8)) * remove step-security/harden-runner from workflows ([#246](#246)) ([c5708d8](c5708d8)) --- This PR was generated with [Release Please](https://github.com/googleapis/release-please). See [documentation](https://github.com/googleapis/release-please#release-please). Co-authored-by: hve-core-release-please[bot] <254602402+hve-core-release-please[bot]@users.noreply.github.com>
…it window issue (#342) ## Description Configure release-please search depths to prevent version detection failures when the latest release tag is beyond the default commit scanning window. This addresses a known limitation ([googleapis/release-please#2267](googleapis/release-please#2267)) where release-please fails to find tags older than ~250 commits, causing it to reprocess the entire commit history and incorrectly detect already-released breaking changes. - Added `release-search-depth: 800` at root level to increase release pagination depth - Added `commit-search-depth: 1000` at root level to increase commit scanning depth - Added `include-component-in-tag: true` to explicitly match the existing tag format (`hve-core-vX.Y.Z`) ## Related Issue(s) Related to [googleapis/release-please#2267](googleapis/release-please#2267) ## Type of Change Select all that apply: **Code & Documentation:** - [x] Bug fix (non-breaking change fixing an issue) - [ ] New feature (non-breaking change adding functionality) - [ ] Breaking change (fix or feature causing existing functionality to change) - [ ] Documentation update **Infrastructure & Configuration:** - [ ] GitHub Actions workflow - [ ] Linting configuration (markdown, PowerShell, etc.) - [ ] Security configuration - [ ] DevContainer configuration - [ ] Dependency update **AI Artifacts:** - [ ] Reviewed contribution with `prompt-builder` agent and addressed all feedback - [ ] Copilot instructions (`.github/instructions/*.instructions.md`) - [ ] Copilot prompt (`.github/prompts/*.prompt.md`) - [ ] Copilot agent (`.github/agents/*.agent.md`) > **Note for AI Artifact Contributors**: > > - **Agents**: Research, indexing/referencing other project (using standard VS Code GitHub Copilot/MCP tools), planning, and general implementation agents likely already exist. Review `.github/agents/` before creating new ones. > - **Model Versions**: Only contributions targeting the **latest Anthropic and OpenAI models** will be accepted. Older model versions (e.g., GPT-3.5, Claude 3) will be rejected. > - See [Agents Not Accepted](../docs/contributing/custom-agents.md#agents-not-accepted) and [Model Version Requirements](../docs/contributing/ai-artifacts-common.md#model-version-requirements). **Other:** - [ ] Script/automation (`.ps1`, `.sh`, `.py`) - [ ] Other (please describe): ## Sample Prompts (for AI Artifact Contributions) N/A - This is a configuration fix, not an AI artifact contribution. ## Testing - Validated JSON syntax using `node -e "require('./release-please-config.json')"` - Verified configuration matches the release-please JSON schema (no IDE validation errors) - Expected behavior: Next release-please run will correctly detect the v2.0.1 tag regardless of commit distance ## Checklist ### Required Checks - [x] Documentation is updated (if applicable) - [x] Files follow existing naming conventions - [x] Changes are backwards compatible (if applicable) - [ ] Tests added for new functionality (if applicable) ### AI Artifact Contributions N/A ### Required Automated Checks The following validation commands must pass before merging: - [ ] Markdown linting: `npm run lint:md` - [ ] Spell checking: `npm run spell-check` - [ ] Frontmatter validation: `npm run lint:frontmatter` - [ ] Link validation: `npm run lint:md-links` - [ ] PowerShell analysis: `npm run lint:ps` ## Security Considerations - [x] This PR does not contain any sensitive or NDA information - [ ] Any new dependencies have been reviewed for security issues - [x] Security-related scripts follow the principle of least privilege ## Additional Notes This fix addresses the erroneous v3.0.0 release PR (#341) that was created immediately after publishing v2.0.1. The root cause was the 250-commit window limitation causing release-please to reprocess the entire commit history and re-detect the breaking change from PR #277 that was already released in v2.0.0. 🔧 - Generated by Copilot
… conversion (#545) ## Description Fixes a race condition where release-please `"draft": true` configuration caused draft releases that are excluded from the GitHub Releases API "latest" query, making them invisible to release-please's own version anchoring within the same workflow invocation. Because the draft release is invisible, release-please scans the full commit history, finds an old breaking change (PR #277), and proposes erroneous v3.0.0 major version bumps. Also added auto-detection of version from the latest GitHub release tag in both extension publish workflows, removing the requirement for manual version input on every dispatch. **Proof the race condition fires in production:** after v2.3.4 merged, release-please created the v2.3.4 release as draft at 22:05:23 UTC. At 22:08:14 UTC — while v2.3.4 was still draft — release-please opened PR #547 proposing v3.0.0. The v2.3.4 release was not published until 22:10:08 UTC, confirming that release-please computed the next version while v2.3.4 was invisible to its own Releases API query. Additionally, v2.3.2 remains stuck in draft state, demonstrating that the `publish-release` job's `gh release edit --draft=false` is unreliable when it depends on upstream jobs that may not always run. ### Changes - fix(workflows): removed `"draft": true` from root and package level in `release-please-config.json` so release-please creates published releases that are immediately visible to the Releases API for version anchoring - fix(workflows): removed `"force-tag-creation": true` from `release-please-config.json` since it requires release-please v17.2.0+ and was silently ignored by the bundled v17.1.3 - fix(workflows): replaced 20-line tag bridge step in `main.yml` with a `gh release edit --draft=true` post-creation conversion, allowing assets to be uploaded to a mutable draft release while preserving release visibility for version anchoring - feat(workflows): added auto-detection of version from latest GitHub release tag in `extension-publish-prerelease.yml` when version input is empty, with ODD minor derivation for the pre-release channel - feat(workflows): added auto-detection of version from latest GitHub release tag in `extension-publish.yml` when version input is empty, with `hve-core-v` prefix stripping - fix(workflows): updated `extension-publish.yml` input description to match new auto-detect behavior (was incorrectly referencing `package.json`) ## Related Issue(s) Fixes #543 ## Type of Change Select all that apply: **Code & Documentation:** - [x] Bug fix (non-breaking change fixing an issue) - [ ] New feature (non-breaking change adding functionality) - [ ] Breaking change (fix or feature causing existing functionality to change) - [ ] Documentation update **Infrastructure & Configuration:** - [x] GitHub Actions workflow - [ ] Linting configuration (markdown, PowerShell, etc.) - [ ] Security configuration - [ ] DevContainer configuration - [ ] Dependency update **AI Artifacts:** - [ ] Reviewed contribution with `prompt-builder` agent and addressed all feedback - [ ] Copilot instructions (`.github/instructions/*.instructions.md`) - [ ] Copilot prompt (`.github/prompts/*.prompt.md`) - [ ] Copilot agent (`.github/agents/*.agent.md`) - [ ] Copilot skill (`.github/skills/*/SKILL.md`) **Other:** - [ ] Script/automation (`.ps1`, `.sh`, `.py`) - [ ] Other (please describe): ## Testing - Validated YAML linting passes via `npm run lint:yaml` — all 25 workflow files passed - Verified `release-please-config.json` is valid JSON with no schema errors - Confirmed the `publish-release` job already contains `gh release edit --draft=false` to finalize releases after asset upload - Confirmed `gh release view --json tagName` correctly returns the latest release tag format (`hve-core-v2.3.4`) ## Checklist ### Required Checks - [x] Documentation is updated (if applicable) - [x] Files follow existing naming conventions - [x] Changes are backwards compatible (if applicable) - [ ] Tests added for new functionality (if applicable) ### Required Automated Checks The following validation commands must pass before merging: - [x] Markdown linting: `npm run lint:md` - [ ] Spell checking: `npm run spell-check` - [ ] Frontmatter validation: `npm run lint:frontmatter` - [ ] Link validation: `npm run lint:md-links` - [ ] PowerShell analysis: `npm run lint:ps` ## Security Considerations - [x] This PR does not contain any sensitive or NDA information - [ ] Any new dependencies have been reviewed for security issues - [x] Security-related scripts follow the principle of least privilege ## Additional Notes The release lifecycle after this change follows a four-phase sequence: release-please creates a **published** release (tag created, visible to Releases API) → post-creation step converts to **draft** (mutable for asset upload) → package/attest/upload jobs attach assets → `publish-release` job converts back to **published**. This preserves the original HTTP 422 fix from PR #538 while eliminating the version anchoring race condition that caused PRs #530, #532, #534, #539, #540, #542, and #547. ### Why the tag bridge didn't work The tag bridge step from PR #538 manually created git tags after draft release creation. However, release-please anchors its version calculation on the GitHub **Releases** API (not the Tags API). Even though tag `hve-core-v2.3.4` existed at commit `9a72f2b`, release-please's query for the latest release excluded the draft, causing it to scan the full commit history and hit the old breaking change from PR #277.
feat(agents): add Task Reviewer and expand RPI to 4-phase workflow
Introduces the Task Reviewer agent and expands the RPI (Research-Plan-Implement) workflow to include a Review phase, creating a complete 4-phase methodology with iteration loops. Also adds the memory agent for conversation context persistence, doc-ops agent for documentation quality assurance, and new supporting prompts for autonomous workflows.
Description
{{YYYY-MM-DD}}syntax across all agents and promptsRPI Updates:
RPI agent now combines Task Researcher, Task Planner, Task Implementor, and Task Reviewer all in one custom agent and utilizes subagents for every step.
/rpi {describe your task you want researched, planned, and implemented}RPI Checkpoint:
When your context is getting too large and you want to continue with a fresh context:
/checkpoint /checkpoint continueNew Task Reviewer:
New Task Reviewer which will review the implemented changes against planned and research work.
Related Issue(s)
N/A
Type of Change
Select all that apply:
Code & Documentation:
Infrastructure & Configuration:
AI Artifacts:
prompt-builderagent and addressed all feedback.github/instructions/*.instructions.md).github/prompts/*.prompt.md).github/agents/*.agent.md)Other:
.ps1,.sh,.py)Sample Prompts (for AI Artifact Contributions)
User Request:
/task-review- Validate implementation against research and plan specifications after completing task implementationExecution Flow:
Output Artifacts:
.copilot-tracking/reviews/{{YYYY-MM-DD}}-task-description-review.mdSuccess Indicators:
For detailed contribution requirements, see:
Testing
Tested task-reviewer agent through complete RPI workflow cycles, validating:
Checklist
Required Checks
AI Artifact Contributions
/prompt-analyzeto review contributionprompt-builderreviewRequired Automated Checks
The following validation commands must pass before merging:
npm run lint:mdnpm run spell-checknpm run lint:frontmatternpm run lint:md-linksnpm run lint:psSecurity Considerations
Additional Notes
This PR expands the RPI methodology from 3 phases to 4 phases by adding the Review phase. The review phase enables:
The memory agent provides session continuity for long-running tasks across conversation resets.
The doc-ops agent operates autonomously using subagents for documentation quality assurance, separate from the RPI workflow.
🤖 - Generated by Copilot