Conversation
…98 #96) PO ownership & scope session (#95 #93): - PO is sole owner of .feature files and all moves (backlog→in-progress→completed) - SE/reviewer never self-select features; escalate via TODO.md System overview & ADR restructure (#98): - Retire docs/architecture.md and docs/c4/ - Create docs/system.md (SE-owned, current system state) - Create docs/context.md and docs/container.md (flat, not in c4/) - Rename docs/scientific-research/ → docs/research/ Reduce agent reads & information architecture (#96): - Rename docs/discovery_journal.md → docs/scope_journal.md - Split docs/discovery.md: PO-owned append-only synthesis + SE-owned docs/domain-model.md - ADR one-file-per-decision: docs/adr/ADR-YYYY-MM-DD-<slug>.md (delete logs.md) - Extract all inline templates to .template files in skill directories - Update all skills with reduced read policies (targeted reads only) - Update self-declaration format: 'As a <role> I declare that:' - Update all research doc references and C4 diagram sources
…e-agnostic patterns (#94) - docs/research/oop-design.md: add entry #36 for refactoring.guru with full smell categories (Bloaters, OO Abusers, Change Preventers, Dispensables, Couplers), ~70 technique categories, and GoF pattern cross-reference - refactor/SKILL.md v2.0: expand smell table from 12 to 22 entries in 5 named categories; replace all Python code examples with language-agnostic trigger/outcome descriptions; replace SOLID Python signals with violation signals; replace Python Zen section with Design Clarity Signals - apply-patterns/SKILL.md v3.0: replace 8 Python before/after code blocks with language-agnostic Before/After structural descriptions and Key structural change callouts; clean Quick Lookup table; add refactoring.guru to sources
- Redesign docs/index.html from 3 cards to 6 curated cards: - System Overview (system.md) - Context Diagram (context.md) - Features (GitHub folder link for backlog/in-progress/completed) - API Reference (api/app.html) - Coverage Report (coverage/index.html) - Research Library (GitHub folder link for 10 papers) - Keep same dark card UI/UX as original - Features and Research link to GitHub for folder browsing - Remove intermediate features/index.html sub-page
Research (entries 59–62 in docs/research/architecture.md): - ATAM: architecture evaluation with scenario analysis - Conway's Law / Inverse Conway Maneuver: team structure mirrors architecture - Architect as Decision-Maker (Fowler + Martin): policy/detail separation - Team Topologies: cognitive load, collaboration, X-as-a-Service Agent changes: - New : owns Step 2 (architecture) and Step 4 (review) - Update : owns Step 3 only (TDD loop, implementation) - Delete : absorbed into system-architect Skill changes: - New : Step 2 architecture protocol (split from implement) - Shrink to Step 3 TDD loop only (v5.0) - Update audience to system-architect (v6.0) - Update read policy for SA/SE split - Update , , , for SA references AGENTS.md updates: - Closed loop: SA designs → SE builds → SA reviews - Update all role descriptions, skills table, filesystem ownership - Update verification philosophy, bug handling, quality gates
- Line 95: SE → SA formalization at Step 2 - Quality Gate: remove architect code-design check (SA Step 4, not SE gate) - Fix numbering skip (1→2→3→5 → 1→2→3) - Fix typo: appropriated → appropriate design patterns - Skills table: clarify create-pr/git-release/update-docs as post-acceptance
README.md: - Step 2 ARCH: Software Engineer → System Architect - Step 4 VERIFY: Reviewer → System Architect - Agent table: @Reviewer → @System-Architect, SE drops Architecture - adversarial review line: reviewer → system-architect Active skills: - define-scope/SKILL.md: SE creates/formalizes domain-model → SA - update-docs/SKILL.md: SE owns system.md/domain-model.md → SA - verify/SKILL.md: code review → semantic review (3 occurrences) - create-pr/SKILL.md: reviewer → system-architect - run-session/SKILL.md: SE Step 2 example → SA + skill architect - select-feature/SKILL.md: SE Step 2 routing → SA + skill architect Research docs: - ai-agents.md: 4 agents (with reviewer) → 5 agents (with SA + designer) - cognitive-science.md: reviewer.md → system-architect.md - testing.md: reviewer.md → verify/SKILL.md - documentation.md: SE updates/rewrites system.md → SA - architecture.md: SE creates ADRs → SA - domain-modeling.md: SE formalizes domain-model → SA
…ilures - Delete 2 old ping-pong-cli post-mortem artifacts (not template docs) - Add docs/post-mortem/.gitkeep to keep empty directory - Add .opencode/skills/define-scope/post-mortem.md.template (max 15 lines) - Add Post-Mortem Protocol section to define-scope/SKILL.md (trigger, workflow, rules) - Update AGENTS.md: Acceptance Failure & Restart section + post-mortem in filesystem tree - Update run-session/SKILL.md: add docs/post-mortem/ to PO read list Workflow: stakeholder failure after acceptance → PO keeps feature in in-progress/ → team writes compact post-mortem → PO scans/selects relevant files → PO reads them → resets TODO.md to Step 2 → SA restarts architecture with post-mortem context.
…ty protocol Add version-control skill: - .opencode/skills/version-control/SKILL.md — Git safety, branch lifecycle, commit hygiene, merge to main with --no-ff, post-mortem branches, conflict detection. Absolute rules: no force push, no history rewrite on pushed branches, use git revert, no commits directly to main. Update AGENTS.md: - Add Branch Model section (normal flow + post-mortem flow + Git Safety Protocol) - Update Step 2/5 descriptions to include branching - Update Acceptance Failure & Restart with fix/<stem> branch from start commit - Update Release Management: releases from main only - Add version-control to skills table Update agents: - software-engineer.md: add version-control to Available Skills and Step Routing - system-architect.md: mention branch verification at Step 2 Update skills: - run-session/SKILL.md: add git branch --show-current check in session start - architect/SKILL.md: add branch verification prerequisite (#5) - implement/SKILL.md: add branch prerequisite, add Branch Hygiene before handoff - verify/SKILL.md: add Branch Gate (section 3), fix section numbering cascade - define-scope/SKILL.md: post-mortem workflow now includes exact git commands for creating fix/<stem> branch from start commit - git-release/SKILL.md: add main-branch guard and ff-only pull - create-pr/SKILL.md: replace squash merge with --no-ff merge - create-skill/SKILL.md: update create-pr description to --no-ff Add research: - docs/research/version-control.md — Pro Git, Driessen git-flow, Git cheat sheet, Fowler branching patterns, merge vs rebase guidance
The system-architect now owns PR creation and merge, consistent with the closed-loop model where SA designs → SE builds → SA reviews → SA integrates. - AGENTS.md: create-pr owner → system-architect - create-pr/SKILL.md: author/audience → system-architect - create-skill/SKILL.md: update skills table reference - software-engineer.md: remove create-pr from Step 5 routing and Available Skills - system-architect.md: add create-pr to Step 5 routing and Available Skills
The stakeholder (human) now owns the release decision. When the stakeholder requests a release, the SA or SE loads the git-release skill to execute it. - AGENTS.md: git-release owner → stakeholder; update release command reference - git-release/SKILL.md: author/audience → stakeholder - create-skill/SKILL.md: update skills table reference - software-engineer.md: remove git-release from Step 5 routing and Available Skills - product-owner.md: update acceptance note to reflect SA creates PR, stakeholder triggers release
Create flow state machine skill: - .opencode/skills/flow/SKILL.md — full state machine with auto-detection rules, prerequisites table, session protocol, and recovery from interruption - .opencode/skills/flow/flow.md.template — idle state template Replace TODO.md with FLOW.md: - FLOW.md in project root — tracks feature, branch, status, session log, next action - Delete TODO.md (superseded) Update all references across 11 files: - run-session/SKILL.md (v5.0): read FLOW.md, detect-state protocol, simplified format - product-owner.md: 6 references updated - software-engineer.md: 4 references updated - system-architect.md: 4 references updated - select-feature/SKILL.md: load when Status is [IDLE], FLOW.md format - define-scope/SKILL.md: reset FLOW.md Status - implement/SKILL.md: build test list, mark in FLOW.md Session Log - verify/SKILL.md: update FLOW.md Next - architect/SKILL.md: 4 references updated - AGENTS.md: session management, escalation, post-mortem reset
…l research Consistency fixes: - AGENTS.md skills table: add skill (was missing) - AGENTS.md filesystem tree: add at root - README.md: add 'No ambiguity on workflow state' bullet mentioning FLOW.md Research additions: - docs/research/ai-agents.md: add 6 entries (72–77) on communication protocols: * Actor Model (Hewitt 1973) — message-passing ownership * CSP (Hoare 1978) — synchronous handoff * Session Types (Honda 1993) — protocol conformance * Statecharts (Harel 1987) — hierarchical states * Design by Contract (Meyer 1986) — preconditions/postconditions * Petri Nets (Petri 1962) — places, transitions, token flow - Update bibliography with all new sources
…, dark stone banner
…cept - logo.svg: transparent-bg temple mark, 4 columns, gold lemniscate in pediment - banner.svg: dark stone bg, TEMPLE·VIII typeset, left mark + right ∞ echo
…subtitle and right echo
…ts, branding documented
…update 8-glyph in assets
…yPI pipeline, 8-glyph assets, FLOW protocol
…-glyph, README, FLOW protocol)
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
temple8across all user-authored files (pyproject.toml, README, template-config.yaml, docs/index.html, setup-project agent, git remote).github/workflows/tag-release.yml(auto-tags on pyproject.toml version bump) and.github/workflows/pypi-publish.yml(build, publish via OIDC, GitHub Release)flowskill with full state machine, auto-detection rules, and templates