mirror of
https://github.com/boshu2/agentops.git
synced 2026-09-14 15:08:13 +08:00
fix(docs): align ASCII box-drawing diagrams across 23 files
Fix right-side │ border alignment in box diagrams. Three fix types: - Trailing space adjustment for border alignment - Box border widening for content that exceeded width - Bottom border dash count matching top border
This commit is contained in:
Boden Fuller
@@ -11,7 +11,7 @@ The architecture rests on five pillars. Each one is independent — you can use
|
||||
```
|
||||
┌─────────────────────────────────────────────────────────────────┐
|
||||
│ DevOps Three Ways │
|
||||
│ (Flow · Feedback · Continual Learning) │
|
||||
│ (Flow · Feedback · Continual Learning) │
|
||||
├─────────────┬─────────────┬─────────────┬───────────────────────┤
|
||||
│ Brownian │ Ralph │ Knowledge │ Fractal │
|
||||
│ Ratchet │ Wiggum │ Flywheel │ Composition │
|
||||
|
||||
+25
-25
@@ -19,40 +19,40 @@ The validate workflow runs **29 jobs** across 4 tiers of parallelism. Most jobs
|
||||
### Job Dependency Graph
|
||||
|
||||
```text
|
||||
┌─────────────────────────────────────────────┐
|
||||
│ 25 independent parallel jobs │
|
||||
│ │
|
||||
│ doc-release-gate smoke-test │
|
||||
┌───────────────────────────────────────────────┐
|
||||
│ 25 independent parallel jobs │
|
||||
│ │
|
||||
│ doc-release-gate smoke-test │
|
||||
│ hook-preflight validate-hooks-doc-parity│
|
||||
│ validate-ci-policy-parity │
|
||||
│ codex-runtime-sections │
|
||||
│ embedded-sync cli-docs-parity │
|
||||
│ shellcheck markdownlint │
|
||||
│ security-scan security-toolchain-gate │
|
||||
│ skill-integrity skill-schema │
|
||||
│ skill-dependency-check │
|
||||
│ validate-ci-policy-parity │
|
||||
│ codex-runtime-sections │
|
||||
│ embedded-sync cli-docs-parity │
|
||||
│ shellcheck markdownlint │
|
||||
│ security-scan security-toolchain-gate │
|
||||
│ skill-integrity skill-schema │
|
||||
│ skill-dependency-check │
|
||||
│ contract-compatibility-gate │
|
||||
│ memrl-health plugin-load-test │
|
||||
│ go-build cli-integration │
|
||||
│ file-manifest-overlap │
|
||||
│ skill-lint learning-coherence │
|
||||
│ bats-tests check-test-staleness │
|
||||
└──────────────┬──────────────────────────────┘
|
||||
│ memrl-health plugin-load-test │
|
||||
│ go-build cli-integration │
|
||||
│ file-manifest-overlap │
|
||||
│ skill-lint learning-coherence │
|
||||
│ bats-tests check-test-staleness │
|
||||
└────────────────┬──────────────────────────────┘
|
||||
│
|
||||
┌──────────────┴──────────────┐
|
||||
│ go-build (must complete) │
|
||||
│ go-build (must complete) │
|
||||
└──┬─────────────┬─────────┬──┘
|
||||
│ │ │
|
||||
┌─────┴──┐ ┌──────┴───┐ ┌───┴──────────┐
|
||||
│ doctor- │ │coverage- │ │json-flag- │
|
||||
│ check │ │ ratchet │ │consistency │
|
||||
└────┬────┘ └────┬─────┘ └──────┬───────┘
|
||||
│ doctor- │ │coverage- │ │json-flag- │
|
||||
│ check │ │ ratchet │ │consistency │
|
||||
└────┬────┘ └────┬─────┘ └──────┬──────┘
|
||||
│ │ │
|
||||
┌─┴────────────┴──────────────┴─┐
|
||||
│ summary │
|
||||
│ (needs: ALL 28 jobs) │
|
||||
│ if: always() │
|
||||
└────────────────────────────────┘
|
||||
│ summary │
|
||||
│ (needs: ALL 28 jobs) │
|
||||
│ if: always() │
|
||||
└───────────────────────────────┘
|
||||
```
|
||||
|
||||
### The `summary` Aggregator Pattern
|
||||
|
||||
+10
-10
@@ -10,21 +10,21 @@ Releases are triggered by git tags and use a publisher-only workflow:
|
||||
git tag vX.Y.Z
|
||||
↓
|
||||
┌──────────────────────────────────────────────────────┐
|
||||
│ Local gate (authoritative) │
|
||||
│ Local gate (authoritative) │
|
||||
├──────────────────────────────────────────────────────┤
|
||||
│ ./scripts/ci-local-release.sh │
|
||||
│ - validation + tests + smoke checks │
|
||||
│ - hook install + ao rpi smoke │
|
||||
│ - SBOM + security report artifacts │
|
||||
│ ./scripts/ci-local-release.sh │
|
||||
│ - validation + tests + smoke checks │
|
||||
│ - hook install + ao rpi smoke │
|
||||
│ - SBOM + security report artifacts │
|
||||
└──────────────────────────────────────────────────────┘
|
||||
↓
|
||||
┌──────────────────────────────────────────────────────┐
|
||||
│ release.yml (publisher only) │
|
||||
│ release.yml (publisher only) │
|
||||
├──────────────────────────────────────────────────────┤
|
||||
│ - GoReleaser publish │
|
||||
│ - GitHub Release notes + assets │
|
||||
│ - Homebrew update │
|
||||
│ - SLSA provenance attestation │
|
||||
│ - GoReleaser publish │
|
||||
│ - GitHub Release notes + assets │
|
||||
│ - Homebrew update │
|
||||
│ - SLSA provenance attestation │
|
||||
└──────────────────────────────────────────────────────┘
|
||||
```
|
||||
|
||||
|
||||
@@ -123,9 +123,9 @@ GOALS.md
|
||||
|
||||
```
|
||||
┌──────────────────────────────┐
|
||||
│ /council │
|
||||
│ (independent reviewers │
|
||||
│ debate, verdict gates work) │
|
||||
│ /council │
|
||||
│ (independent reviewers │
|
||||
│ debate, verdict gates work)│
|
||||
└───────────┬──────────────────┘
|
||||
│ used by
|
||||
┌────────────────────┼────────────────────┐
|
||||
|
||||
@@ -24,7 +24,7 @@ The top-level skill `/rpi` chains the full pipeline. Each node is a skill. Arrow
|
||||
┌───────────────────────────────────────────────────────────────────┐
|
||||
│ /rpi │
|
||||
│ (full pipeline orchestrator) │
|
||||
└──┬──────────┬───────────┬─────────────┬──────────┬───────────────┘
|
||||
└──┬──────────┬───────────┬─────────────┬──────────┬─────────────── ┘
|
||||
│ │ │ │ │
|
||||
▼ ▼ ▼ ▼ ▼
|
||||
/research /plan /pre-mortem /crank /post-mortem
|
||||
@@ -33,9 +33,9 @@ The top-level skill `/rpi` chains the full pipeline. Each node is a skill. Arrow
|
||||
│ │ │ └── calls /retro
|
||||
│ │ │
|
||||
│ │ ┌──────────┴──────────┐
|
||||
│ │ │ /crank │
|
||||
│ │ │ (wave executor) │
|
||||
│ │ └──────────┬───────────┘
|
||||
│ │ │ /crank │
|
||||
│ │ │ (wave executor) │
|
||||
│ │ └──────────┬──────────┘
|
||||
│ │ │ spawns N parallel
|
||||
│ │ ▼
|
||||
│ │ /implement
|
||||
@@ -58,9 +58,9 @@ The top-level skill `/rpi` chains the full pipeline. Each node is a skill. Arrow
|
||||
|
||||
```
|
||||
┌──────────────────────────────┐
|
||||
│ /council │
|
||||
│ (independent judges debate, │
|
||||
│ verdict gates delivery) │
|
||||
│ /council │
|
||||
│ (independent judges debate, │
|
||||
│ verdict gates delivery) │
|
||||
└───────────┬──────────────────┘
|
||||
│ used by
|
||||
┌────────────────────┼────────────────────┐
|
||||
|
||||
@@ -12,14 +12,14 @@ PDC (Prevent, Detect, Correct) is a three-phase framework for maintaining qualit
|
||||
|
||||
```
|
||||
┌──────────────────────────────────────────────────────────┐
|
||||
│ PREVENT │
|
||||
│ Stop problems before they enter the system │
|
||||
│ PREVENT │
|
||||
│ Stop problems before they enter the system │
|
||||
├──────────────────────────────────────────────────────────┤
|
||||
│ DETECT │
|
||||
│ Catch problems as soon as they occur │
|
||||
│ DETECT │
|
||||
│ Catch problems as soon as they occur │
|
||||
├──────────────────────────────────────────────────────────┤
|
||||
│ CORRECT │
|
||||
│ Fix problems and prevent recurrence │
|
||||
│ CORRECT │
|
||||
│ Fix problems and prevent recurrence │
|
||||
└──────────────────────────────────────────────────────────┘
|
||||
```
|
||||
|
||||
|
||||
@@ -81,12 +81,12 @@ FIRE is the reconciliation engine that implements the Brownian Ratchet:
|
||||
|
||||
```
|
||||
┌──────────────────────────────────────────────────────────────┐
|
||||
│ FIRE LOOP │
|
||||
│ │
|
||||
│ FIRE LOOP │
|
||||
│ │
|
||||
│ FIND ────► IGNITE ────► REAP ────► ESCALATE │
|
||||
│ (state) (chaos) (ratchet) (recovery) │
|
||||
│ │ │ │
|
||||
│ └───────────────────────────────────┘ │
|
||||
│ │ │ │
|
||||
│ └───────────────────────────────────┘ │
|
||||
│ (loop) │
|
||||
│ │
|
||||
│ EXIT when: all work reaped │
|
||||
@@ -364,8 +364,8 @@ FIRE Loop (Ratchet) Knowledge Flywheel
|
||||
│ │
|
||||
▼ ▼
|
||||
┌─────────────────────────────────────────────────┐
|
||||
│ NEXT FIRE CYCLE │
|
||||
│ (with better knowledge = faster/cheaper) │
|
||||
│ NEXT FIRE CYCLE │
|
||||
│ (with better knowledge = faster/cheaper) │
|
||||
└─────────────────────────────────────────────────┘
|
||||
```
|
||||
|
||||
|
||||
@@ -26,11 +26,11 @@ The Claude Code plugin ecosystem has exploded. Here's how the major players stac
|
||||
│ WHAT OTHERS OPTIMIZE WHAT AGENTOPS OPTIMIZES │
|
||||
│ ══════════════════════ ═════════════════════════ │
|
||||
│ │
|
||||
│ Session 1 Session 2 Session 3 Session 1 Session 2 Session 3│
|
||||
│ ┌──────┐ ┌──────┐ ┌──────┐ ┌──────┐ ┌──────┐ ┌──────┐ │
|
||||
│ │ Fast │ │ Fast │ │ Fast │ │Learn │ → │Recall│ → │Expert│ │
|
||||
│ │ │ │ │ │ │ │ │ │ │ │ │ │
|
||||
│ └──────┘ └──────┘ └──────┘ └──────┘ └──────┘ └──────┘ │
|
||||
│ Session 1 Session 2 Session 3 Session 1 Session 2 Session 3 │
|
||||
│ ┌──────┐ ┌──────┐ ┌──────┐ ┌──────┐ ┌──────┐ ┌──────┐ │
|
||||
│ │ Fast │ │ Fast │ │ Fast │ │Learn │ → │Recall│ → │Expert│ │
|
||||
│ │ │ │ │ │ │ │ │ │ │ │ │ │
|
||||
│ └──────┘ └──────┘ └──────┘ └──────┘ └──────┘ └──────┘ │
|
||||
│ ↓ ↓ ↓ │ │ │ │
|
||||
│ [reset] [reset] [reset] └──────────┴──────────┘ │
|
||||
│ COMPOUNDS │
|
||||
|
||||
@@ -54,9 +54,9 @@ V3 introduced swarms that can adapt their behavior within a session.
|
||||
┌─────────────────────────────────────────────────────────────────┐
|
||||
│ CLAUDE-FLOW │
|
||||
│ │
|
||||
│ Session 1: 60 agents solve auth bug │
|
||||
│ Session 2: 60 agents solve auth bug (no memory of Session 1) │
|
||||
│ Session 3: 60 agents solve auth bug (no memory of Session 2) │
|
||||
│ Session 1: 60 agents solve auth bug │
|
||||
│ Session 2: 60 agents solve auth bug (no memory of Session 1) │
|
||||
│ Session 3: 60 agents solve auth bug (no memory of Session 2) │
|
||||
│ │
|
||||
│ Fast parallel execution, but no compounding │
|
||||
└─────────────────────────────────────────────────────────────────┘
|
||||
@@ -64,8 +64,8 @@ V3 introduced swarms that can adapt their behavior within a session.
|
||||
┌─────────────────────────────────────────────────────────────────┐
|
||||
│ AGENTOPS │
|
||||
│ │
|
||||
│ Session 1: Solve auth bug → Extract: "token refresh" │
|
||||
│ Session 2: Auth issue? ← Inject prior knowledge │
|
||||
│ Session 1: Solve auth bug → Extract: "token refresh" │
|
||||
│ Session 2: Auth issue? ← Inject prior knowledge │
|
||||
│ Session 3: Auth? ← Instant domain expertise │
|
||||
│ │
|
||||
│ Knowledge persists and compounds │
|
||||
@@ -125,9 +125,9 @@ Claude-Flow excels at coordinating many agents. It doesn't provide the deep sema
|
||||
┌─────────────────────────────────────────────────────────────────┐
|
||||
│ CLAUDE-FLOW V3 │
|
||||
│ │
|
||||
│ ┌─────────┐ ┌─────────────────────────────────────────┐ │
|
||||
│ ┌─────────┐ ┌─────────────────────────────────────────┐ │
|
||||
│ │ Task │───▶│ SWARM COORDINATOR │ │
|
||||
│ └─────────┘ └─────────────────────────────────────────┘ │
|
||||
│ └─────────┘ └─────────────────────────────────────────┘ │
|
||||
│ │ │
|
||||
│ ┌────────────────────┼────────────────────┐ │
|
||||
│ ▼ ▼ ▼ │
|
||||
@@ -220,16 +220,16 @@ AgentOps: ████████████░░░░░░░░
|
||||
│ CLAUDE-FLOW + AGENTOPS │
|
||||
│ │
|
||||
│ SessionStart: │
|
||||
│ └── AgentOps injects prior knowledge │
|
||||
│ └── AgentOps injects prior knowledge │
|
||||
│ │
|
||||
│ Execution: │
|
||||
│ └── Claude-Flow orchestrates 60+ agents │
|
||||
│ └── Claude-Flow orchestrates 60+ agents │
|
||||
│ │
|
||||
│ Validation: │
|
||||
│ └── AgentOps /vibe validates all outputs │
|
||||
│ └── AgentOps /vibe validates all outputs │
|
||||
│ │
|
||||
│ SessionEnd: │
|
||||
│ └── AgentOps extracts learnings for next time │
|
||||
│ └── AgentOps extracts learnings for next time │
|
||||
│ │
|
||||
└─────────────────────────────────────────────────────────────────┘
|
||||
```
|
||||
|
||||
+17
-17
@@ -75,31 +75,31 @@ GSD has persistence (state files, handoffs, threads) but no knowledge flywheel.
|
||||
|
||||
```
|
||||
┌─────────────────────────────────────────────────────────────────┐
|
||||
│ GSD │
|
||||
│ │
|
||||
│ GSD │
|
||||
│ │
|
||||
│ Session 1: discuss → plan → execute → verify → Done │
|
||||
│ ↓ │
|
||||
│ (state saved) │
|
||||
│ ↓ │
|
||||
│ ↓ │
|
||||
│ (state saved) │
|
||||
│ ↓ │
|
||||
│ Session 2: resume-work → (same state, same intelligence) │
|
||||
│ │
|
||||
│ │
|
||||
│ Session 100: (agents are no smarter than session 1) │
|
||||
│ │
|
||||
│ │
|
||||
└─────────────────────────────────────────────────────────────────┘
|
||||
|
||||
┌─────────────────────────────────────────────────────────────────┐
|
||||
│ AGENTOPS │
|
||||
│ │
|
||||
│ Session 1: research → plan → pre-mortem → crank → vibe → retro│
|
||||
│ ↓ │
|
||||
│ (learnings extracted) │
|
||||
│ ↓ │
|
||||
│ (scored and stored) │
|
||||
│ │
|
||||
│ AGENTOPS │
|
||||
│ │
|
||||
│ Session 1: research → plan → pre-mortem → crank → vibe → retro │
|
||||
│ ↓ │
|
||||
│ (learnings extracted) │
|
||||
│ ↓ │
|
||||
│ (scored and stored) │
|
||||
│ │
|
||||
│ Session 2: (inject prior knowledge) → better starting point │
|
||||
│ │
|
||||
│ │
|
||||
│ Session 100: (agent is a domain expert with scored knowledge) │
|
||||
│ │
|
||||
│ │
|
||||
└─────────────────────────────────────────────────────────────────┘
|
||||
```
|
||||
|
||||
|
||||
@@ -249,16 +249,16 @@ AgentOps doesn't aim for "spec-as-source" — it captures *learnings*, not just
|
||||
│ └── Requirements → Design → Tasks (structured specs) │
|
||||
│ │
|
||||
│ AgentOps handles: │
|
||||
│ └── Pre-mortem (failure simulation) │
|
||||
│ └── /vibe (8-aspect validation) │
|
||||
│ └── /post-mortem (learning extraction) │
|
||||
│ └── Cross-session memory │
|
||||
│ └── Pre-mortem (failure simulation) │
|
||||
│ └── /vibe (8-aspect validation) │
|
||||
│ └── /post-mortem (learning extraction) │
|
||||
│ └── Cross-session memory │
|
||||
│ │
|
||||
│ Combined flow: │
|
||||
│ SDD specs → AgentOps pre-mortem → Implement → AgentOps vibe │
|
||||
│ ↓ │
|
||||
│ SDD specs → AgentOps pre-mortem → Implement → AgentOps vibe │
|
||||
│ ↓ │
|
||||
│ Extract learnings │
|
||||
│ ↓ │
|
||||
│ ↓ │
|
||||
│ Next session │
|
||||
└─────────────────────────────────────────────────────────────────┘
|
||||
```
|
||||
|
||||
@@ -46,21 +46,21 @@ Built-in enforcement of software engineering best practices. No over-engineering
|
||||
┌─────────────────────────────────────────────────────────────────┐
|
||||
│ SUPERPOWERS │
|
||||
│ │
|
||||
│ Session 1: Debug auth bug [learned: token refresh] │
|
||||
│ Session 2: Debug auth bug [learned: token refresh] │
|
||||
│ Session 3: Debug auth bug [learned: token refresh] │
|
||||
│ ↑ │
|
||||
│ Session 1: Debug auth bug [learned: token refresh] │
|
||||
│ Session 2: Debug auth bug [learned: token refresh] │
|
||||
│ Session 3: Debug auth bug [learned: token refresh] │
|
||||
│ ↑ │
|
||||
│ Same learning, every time │
|
||||
└─────────────────────────────────────────────────────────────────┘
|
||||
|
||||
┌─────────────────────────────────────────────────────────────────┐
|
||||
│ AGENTOPS │
|
||||
│ │
|
||||
│ Session 1: Debug auth bug [learned: token refresh] │
|
||||
│ ↓ (stored) │
|
||||
│ Session 2: Auth issue? "I remember this pattern" │
|
||||
│ ↓ (reinforced) │
|
||||
│ Session 3: Auth? *instant recall* │
|
||||
│ Session 1: Debug auth bug [learned: token refresh] │
|
||||
│ ↓ (stored) │
|
||||
│ Session 2: Auth issue? "I remember this pattern" │
|
||||
│ ↓ (reinforced) │
|
||||
│ Session 3: Auth? *instant recall* │
|
||||
│ │
|
||||
└─────────────────────────────────────────────────────────────────┘
|
||||
```
|
||||
|
||||
@@ -18,15 +18,15 @@ The packet has five sections, each with a defined character budget, content sour
|
||||
|
||||
```
|
||||
┌──────────────────────────────────────────────────────────┐
|
||||
│ CONTEXT PACKET │
|
||||
│ │
|
||||
│ ┌─────────┐ ┌─────────┐ ┌─────────┐ │
|
||||
│ CONTEXT PACKET │
|
||||
│ │
|
||||
│ ┌─────────┐ ┌─────────┐ ┌─────────┐ │
|
||||
│ │ GOALS │ │ HISTORY │ │ INTEL │ │
|
||||
│ │ ~2K ch │ │ ~8K ch │ │ ~12K ch │ │
|
||||
│ └─────────┘ └─────────┘ └─────────┘ │
|
||||
│ ┌─────────┐ ┌──────────┐ │
|
||||
│ │ TASK │ │ PROTOCOL │ │
|
||||
│ │ ~4K ch │ │ ~2K ch │ │
|
||||
│ │ TASK │ │ PROTOCOL │ │
|
||||
│ │ ~4K ch │ │ ~2K ch │ │
|
||||
│ └─────────┘ └──────────┘ │
|
||||
│ │
|
||||
│ Total: ~28K chars / ~7K tokens │
|
||||
|
||||
+16
-16
@@ -486,42 +486,42 @@ The pipeline degrades gracefully. Each version is strictly better than the previ
|
||||
Agent Session
|
||||
│
|
||||
▼
|
||||
┌─────────┐ ┌─────────┐
|
||||
┌──────────┐ ┌─────────┐
|
||||
│ /forge │ │ /retro │ (existing skills, unchanged)
|
||||
└────┬─────┘ └────┬────┘
|
||||
│ │
|
||||
▼ ▼
|
||||
┌──────────────────────────┐
|
||||
│ STAGE 1: CATALOG │ v1
|
||||
│ Structure into typed │
|
||||
│ artifacts with metadata │
|
||||
│ → .agents/pool/pending/ │
|
||||
│ Structure into typed │
|
||||
│ artifacts with metadata │
|
||||
│ → .agents/pool/pending/ │
|
||||
└────────────┬─────────────┘
|
||||
│
|
||||
▼
|
||||
┌──────────────────────────┐
|
||||
│ STAGE 2: VERIFY │ v1
|
||||
│ Mechanical checks: │
|
||||
│ tests pass? goals up? │
|
||||
│ → verification metadata │
|
||||
│ Mechanical checks: │
|
||||
│ tests pass? goals up? │
|
||||
│ → verification metadata │
|
||||
└────────────┬─────────────┘
|
||||
│
|
||||
▼
|
||||
┌──────────────────────────┐
|
||||
│ STAGE 3: INDEX │ v2
|
||||
│ Tag by topic, skill, │
|
||||
│ goal. Update search idx │
|
||||
│ → ao search, ao lookup │
|
||||
│ Tag by topic, skill, │
|
||||
│ goal. Update search idx │
|
||||
│ → ao search, ao lookup │
|
||||
└────────────┬─────────────┘
|
||||
│
|
||||
▼
|
||||
┌──────────────────────────┐
|
||||
│ STAGE 4: SCORE │ v2
|
||||
│ 5-dim quality gate: │
|
||||
│ specificity, action- │
|
||||
│ ability, novelty, │
|
||||
│ context, confidence │
|
||||
│ → gold/silver/bronze │
|
||||
│ 5-dim quality gate: │
|
||||
│ specificity, action- │
|
||||
│ ability, novelty, │
|
||||
│ context, confidence │
|
||||
│ → gold/silver/bronze │
|
||||
└────────────┬─────────────┘
|
||||
│
|
||||
┌────┴────┐
|
||||
@@ -540,7 +540,7 @@ The pipeline degrades gracefully. Each version is strictly better than the previ
|
||||
│ CONSTRAIN │
|
||||
│ Compile into │
|
||||
│ hook rules │
|
||||
│ → constraints/│
|
||||
│→ constraints/│
|
||||
└──────────────┘
|
||||
│
|
||||
▼
|
||||
|
||||
@@ -158,7 +158,7 @@ Multi-Session │ Olympus
|
||||
│ │
|
||||
│ ┌────┴────┐
|
||||
│ │ AgentOps │ ← WE ARE HERE
|
||||
│ │(Shift-L) │
|
||||
│ │(Shift-L)│
|
||||
│ └────┬────┘
|
||||
│ │
|
||||
Single-Session │ Agent SDKs
|
||||
|
||||
@@ -347,8 +347,8 @@ profile:
|
||||
┌─────────────────────────────────────────┐
|
||||
│ Workspace Profiles (Production) │ ← Organize reality
|
||||
│ - 5 roles with 100+ real artifacts │
|
||||
│ - Multi-repo coordination │
|
||||
│ - Token budgets for composition │
|
||||
│ - Multi-repo coordination │
|
||||
│ - Token budgets for composition │
|
||||
└─────────────────────────────────────────┘
|
||||
```
|
||||
|
||||
|
||||
+9
-9
@@ -408,10 +408,10 @@ This is why progress can be made one-way at the artifact level, while system-lev
|
||||
|
||||
```
|
||||
┌─────────────────────────────────────────────────────────────────┐
|
||||
│ THE AGENTOPS SYSTEM │
|
||||
│ THE AGENTOPS SYSTEM │
|
||||
├─────────────────────────────────────────────────────────────────┤
|
||||
│ │
|
||||
│ DEVOPS LAYER (The Three Ways) │
|
||||
│ │
|
||||
│ DEVOPS LAYER (The Three Ways) │
|
||||
│ ┌─────────────────────────────────────────────────────────┐ │
|
||||
│ │ Flow: Session → Forge → Store → Inject → Session │ │
|
||||
│ │ Feedback: Vibe, Pre-mortem, Citations, Utility scores │ │
|
||||
@@ -445,13 +445,13 @@ This is why progress can be made one-way at the artifact level, while system-lev
|
||||
│ ▼ │
|
||||
│ THE GOAL (Escape Velocity) │
|
||||
│ ┌─────────────────────────────────────────────────────────┐ │
|
||||
│ │ │ │
|
||||
│ │ σ × ρ > δ │ │
|
||||
│ │ │ │
|
||||
│ │ │ │
|
||||
│ │ σ × ρ > δ │ │
|
||||
│ │ │ │
|
||||
│ │ retrieval × usage > decay │ │
|
||||
│ │ │ │
|
||||
│ │ When true: KNOWLEDGE COMPOUNDS │ │
|
||||
│ │ │ │
|
||||
│ │ │ │
|
||||
│ │ When true: KNOWLEDGE COMPOUNDS │ │
|
||||
│ │ │ │
|
||||
│ └─────────────────────────────────────────────────────────┘ │
|
||||
│ │
|
||||
└─────────────────────────────────────────────────────────────────┘
|
||||
|
||||
@@ -72,7 +72,7 @@ with `ao codex ensure-stop`.
|
||||
|
||||
```
|
||||
┌─────────────────────────────────────────────────────────────┐
|
||||
│ SESSION START │
|
||||
│ SESSION START │
|
||||
│ "Continue the API work" or /session-start │
|
||||
├─────────────────────────────────────────────────────────────┤
|
||||
│ • Load bundle (if exists) │
|
||||
@@ -83,22 +83,22 @@ with `ao codex ensure-stop`.
|
||||
│
|
||||
▼
|
||||
┌─────────────────────────────────────────────────────────────┐
|
||||
│ WORK │
|
||||
│ WORK │
|
||||
│ "Add the validation layer" or just start coding │
|
||||
├─────────────────────────────────────────────────────────────┤
|
||||
│ • Implement features │
|
||||
│ • Fix bugs │
|
||||
│ • Research when needed │
|
||||
│ • Commit frequently │
|
||||
│ │
|
||||
│ Track progress: │
|
||||
│ • Implement features │
|
||||
│ • Fix bugs │
|
||||
│ • Research when needed │
|
||||
│ • Commit frequently │
|
||||
│ │
|
||||
│ Track progress: │
|
||||
│ • "Feature X is done" or /progress-update --complete X │
|
||||
│ • "I'm blocked on Y" or /progress-update --blocker "Y" │
|
||||
└─────────────────────────────────────────────────────────────┘
|
||||
│
|
||||
▼
|
||||
┌─────────────────────────────────────────────────────────────┐
|
||||
│ SESSION END │
|
||||
│ SESSION END │
|
||||
│ "Done for today" or /session-end │
|
||||
├─────────────────────────────────────────────────────────────┤
|
||||
│ • Check for uncommitted changes │
|
||||
@@ -113,7 +113,7 @@ with `ao codex ensure-stop`.
|
||||
|
||||
```
|
||||
┌─────────────────────────────────────────────────────────────┐
|
||||
│ CODEX SESSION START │
|
||||
│ CODEX SESSION START │
|
||||
│ ao codex start │
|
||||
├─────────────────────────────────────────────────────────────┤
|
||||
│ • Inspect .agents/ and surfaced learnings │
|
||||
@@ -134,7 +134,7 @@ with `ao codex ensure-stop`.
|
||||
│
|
||||
▼
|
||||
┌─────────────────────────────────────────────────────────────┐
|
||||
│ CODEX SESSION END │
|
||||
│ CODEX SESSION END │
|
||||
│ ao codex stop │
|
||||
├─────────────────────────────────────────────────────────────┤
|
||||
│ • Resolve archived transcript or history fallback │
|
||||
|
||||
@@ -132,7 +132,7 @@ reads each judge's output file sequentially with the Read tool and synthesizes.
|
||||
```
|
||||
┌─────────────────────────────────────────────────────────────────┐
|
||||
│ Phase 1: Build Packet (JSON) │
|
||||
│ - Task type (validate/brainstorm/research) │
|
||||
│ - Task type (validate/brainstorm/research) │
|
||||
│ - Target description │
|
||||
│ - Context (files, diffs, prior decisions) │
|
||||
│ - Perspectives to assign │
|
||||
@@ -140,9 +140,9 @@ reads each judge's output file sequentially with the Read tool and synthesizes.
|
||||
│
|
||||
▼
|
||||
┌─────────────────────────────────────────────────────────────────┐
|
||||
│ Phase 1a: Select spawn backend │
|
||||
│ codex_subagents | claude_teams | background_fallback │
|
||||
│ Team lead = spawner (this agent) │
|
||||
│ Phase 1a: Select spawn backend │
|
||||
│ codex_subagents | claude_teams | background_fallback │
|
||||
│ Team lead = spawner (this agent) │
|
||||
└─────────────────────────────────────────────────────────────────┘
|
||||
│
|
||||
┌─────────────────┴─────────────────┐
|
||||
|
||||
@@ -10,12 +10,12 @@ FIRE is the reconciliation loop that extracts progress from chaos. Like a forge
|
||||
|
||||
```
|
||||
┌──────────────────────────────────────────────────────────┐
|
||||
│ FIRE LOOP │
|
||||
│ │
|
||||
│ FIRE LOOP │
|
||||
│ │
|
||||
│ FIND ────► IGNITE ────► REAP ────► ESCALATE │
|
||||
│ (state) (chaos) (ratchet) (recovery) │
|
||||
│ │ │ │
|
||||
│ └───────────────────────────────────┘ │
|
||||
│ │ │ │
|
||||
│ └───────────────────────────────────┘ │
|
||||
│ (loop) │
|
||||
│ │
|
||||
│ EXIT when: all children closed │
|
||||
|
||||
@@ -26,7 +26,7 @@ Pick what you need. Every skill works standalone. `/swarm` multiplies any of the
|
||||
│ /implement ────► research + plan + build + vibe│
|
||||
│ /vibe ─────────► /complexity + /council │
|
||||
│ /pre-mortem ───► /council (failure simulation) │
|
||||
│ /post-mortem ──► /council + knowledge lifecycle │
|
||||
│ /post-mortem ──► /council + knowledge lifecycle│
|
||||
│ /council ──────► parallel judges (multi-model) │
|
||||
└─────────────────────────────────────────────────┘
|
||||
```
|
||||
|
||||
@@ -18,25 +18,25 @@
|
||||
## Agent Autonomy Boundaries
|
||||
|
||||
```
|
||||
┌─────────────────────────────────────────────────────┐
|
||||
│ AGENT-AUTONOMOUS (L0–L3) │
|
||||
│ Agent writes tests AND implementation. │
|
||||
│ No human input needed for test design. │
|
||||
│ │
|
||||
│ L0: Contract — from SPEC.md alone │
|
||||
│ L1: Unit — TDD RED→GREEN from spec │
|
||||
│ L2: Integration — from subsystem spec + adapters │
|
||||
┌───────────────────────────────────────────────────────┐
|
||||
│ AGENT-AUTONOMOUS (L0–L3) │
|
||||
│ Agent writes tests AND implementation. │
|
||||
│ No human input needed for test design. │
|
||||
│ │
|
||||
│ L0: Contract — from SPEC.md alone │
|
||||
│ L1: Unit — TDD RED→GREEN from spec │
|
||||
│ L2: Integration — from subsystem spec + adapters │
|
||||
│ L3: Component — agent writes, human defines scenarios│
|
||||
├─────────────────────────────────────────────────────┤
|
||||
│ HUMAN-GUIDED (L4–L7) │
|
||||
│ Human defines WHAT to test. │
|
||||
│ Agent builds the test infrastructure. │
|
||||
│ │
|
||||
│ L4: Smoke — human defines "critical path" │
|
||||
│ L5: E2E — human designs flow, agent harness │
|
||||
│ L6: Acceptance— human only validates │
|
||||
│ L7: Prod — monitors + human judgment │
|
||||
└─────────────────────────────────────────────────────┘
|
||||
├───────────────────────────────────────────────────────┤
|
||||
│ HUMAN-GUIDED (L4–L7) │
|
||||
│ Human defines WHAT to test. │
|
||||
│ Agent builds the test infrastructure. │
|
||||
│ │
|
||||
│ L4: Smoke — human defines "critical path" │
|
||||
│ L5: E2E — human designs flow, agent harness │
|
||||
│ L6: Acceptance— human only validates │
|
||||
│ L7: Prod — monitors + human judgment │
|
||||
└───────────────────────────────────────────────────────┘
|
||||
```
|
||||
|
||||
## RPI Phase Mapping
|
||||
|
||||
@@ -175,10 +175,10 @@ File Ownership Map (Wave N):
|
||||
┌─────────────────────────────┬──────────┬──────────┐
|
||||
│ File │ Owner │ Conflict │
|
||||
├─────────────────────────────┼──────────┼──────────┤
|
||||
│ src/auth/middleware.go │ task-1 │ │
|
||||
│ src/auth/middleware_test.go │ task-1 │ │
|
||||
│ src/api/routes.go │ task-2 │ │
|
||||
│ src/config/settings.go │ task-1,3 │ YES │
|
||||
│ src/auth/middleware.go │ task-1 │ │
|
||||
│ src/auth/middleware_test.go │ task-1 │ │
|
||||
│ src/api/routes.go │ task-2 │ │
|
||||
│ src/config/settings.go │ task-1,3 │ YES │
|
||||
└─────────────────────────────┴──────────┴──────────┘
|
||||
Conflicts: 1 (resolved: serialized task-3 into sub-wave 2)
|
||||
```
|
||||
|
||||
Reference in New Issue
Block a user