2026-07-17
Uncategorized
2026-07-16
Documentation
- #451 add Phase 12b Operator Runbook to /implement-ticket
- #450 defer sections 03 & 05 detail to reference files (~15k chars / ~3.9k tokens always-on reduction)
- #447 add story-size preflight and token-reduction callout to /ticket-triage
- #441 add context-size preflight to /implement-ticket
Uncategorized
2026-07-15
Documentation
Maintenance
2026-07-14
Documentation
2026-07-10
Features
Uncategorized
2026-07-09
Features
- #436 required North Star alignment field on protocol-change RFC template (Layer E)
- #435 advisory vision-alignment-check CI workflow (Layer D)
- #434 North Star alignment section in PR template + CONTRIBUTING (Layer C)
- #433 vision-alignment check in /update-agentic-engineering + Skeptic step 3.6 (Layers A+B)
Documentation
2026-07-08
Features
- #417 cross-harness team health section
- #416 profile-aware installers + install-profiles.sh
- #415 per-role reviewer models (reviewers.by_role)
Fixes
- #432 enable by default across all adapters and templates (DS-47)
- #413 resolve per-role model from team.yml at dispatch
Documentation
- #430 define absent-key defaults for all under-specified config settings
- #428 backfill curated private-store learnings (DS-73 remainder)
Uncategorized
2026-07-07
Uncategorized
- #427 DS-73: correct MEMORY.md injection claim + wire @MEMORY.md import and learnings retrieval
- #426 DS-75: Open-goal loop mandatory caps + minimal wiring; fix batch stall semantics
- #425 DS-77: Tier-3 escalation for authoring roles on ADR-tier units
- #424 DS-76: .omp adapter ships assembled METHODOLOGY.md (fix or label PARTIAL)
2026-07-06
Features
Documentation
2026-07-03
Fixes
Maintenance
Uncategorized
2026-07-02
Features
- #404 /agentic-config - interactive in-session settings command
- #391 DS-62 tracker-wide deterministic sweep + /wrap writeback
- #387 co-locate anti-rationalization backstops in the risk/delegation tables (DS-60)
Fixes
- #405 align detection attribution to wrap-time path across sibling docs
- #399 correct skill-candidate trigger wording + phantom command; rebuild adapters
- #396 resolve real git hooks dir so pre-commit install/uninstall work from a worktree (DS-58)
- #393 make skill-auto-load hook resolve to a real script + resolution test (DS-57)
- #392 update version-check notice extraction for PR #322 phrasing (DS-59)
Documentation
- #410 DS-56 feedback capture + _lib PATH-symlink gotcha
- #407 'just ask your session agent' callouts on config cards
- #402 add Risk Classification (17) and Cross-Session Resume (18) decks
- #400 add Learnings Pipeline (15) and Worktree Lifecycle (16) decks
- #398 add trigger-catalog, graphify, and events-log/telemetry pages
- #389 record DS-48 shipped + fix stale preset example
- #383 sync adapter/harness counts for opencode+copilot worker-set expansion (Unit 6)
- #382 tighten conductor contract for harness-neutral dispatch
- #380 list configure-team in Pi adapter command inventory
Uncategorized
- #411 DS-66: resolve _lib and sibling paths through PATH symlinks (5 bins)
- #409 DS-56: add /feedback-triage to command indexes
- #408 DS-56: wrap Part D.5 feedback capture + /feedback-triage command
- #406 DS-56: resolve _lib through PATH symlink for agentic-feedback
- #403 DS-56: add bin/agentic-feedback home-dir feedback store CLI
- #401 DS-64: goal-condition-evaluator - cheap per-turn stop-condition check for open-goal loops (Tier 1)
- #395 least-privilege permissions on read-only CI workflows (CodeQL #60-66)
- #388 DS-48 Wave 2: bounded multi-file Low override (relaxed) + downward tie-break counterweight
- #376 DS-48 Lever 9 Unit A: remove session-wide preset layer from methodology + docs
2026-07-01
Features
- #377 snapshot hooks to a session-stable dir so a git pull can't rewire live sessions (DS-54 Tier 2, Unit 1)
- #373 warn when a git pull changes hooks/ (DS-54)
- #367 universality (works-for-everyone) North Star pillar + scope /ticket-triage no-args to configured project/team
Documentation
- #378 repurpose hooks-changed advisory for the snapshot model + Step 4.5 refresh (DS-54 Tier 2, Units 2-3)
- #372 per-setting config cards with descriptions (replaces table)
- #371 lead Configuration Layer with a settings-at-a-glance table
- #370 add configuration-reference quick-reference (every setting + default)
- #369 sync #profiles with preset alias + token-spend note
- #368 add token-spend / right-sizing section + preset alias
Uncategorized
2026-06-30
Features
- #366 reliable /wrap lock-wait via agentic-wrap-acquire-lock binary
- #365 implement-ticket delegates >=2-ticket batch planning to /ticket-triage
- #364 add /ticket-triage batch planner (standalone)
- #360 ticket-driven-by-default flow when a tracker is connected
- #358 telemetry-health counter for silent-fail visibility (#266)
- #336 add .telemetry-health.json failure counter + agentic-status surface (#266) [hooks-stack 3/3]
Fixes
Performance
Refactoring
Documentation
- #363 update command and reference doc counts
- #362 add capability-preflight and wrap-deferred end-user pages
- #361 refresh stale What's-included inventory (30 refs, 22 commands)
- #359 add end-user docs page for skill-candidate detection
Maintenance
2026-06-29
Features
- #357 add VS Code Copilot adapter
- #355 add learnings_candidate[] digest field for subagent-internal capture (DS-40)
Fixes
Documentation
2026-06-28
Documentation
2026-06-27
Features
- #351 deterministic events.jsonl emission for ad-hoc sessions
- #350 wire wrap + wrap-ticket to deep-cluster (wrap-time signal)
- #349 enforce-tier PreToolUse hook - mechanical Tier-3 review-escalation backstop
- #348 wrap-time deep-cluster helper (library + CLI)
- #346 /skill-candidates command, session-start notice, scaffolding + canonical entry format
- #344 Layer-2 opt-in in-session nudge
- #343 add skill_candidate_detection + skill_candidate_nudge config toggles
- #342 wire detector into stop-context (session-end scan)
- #341 add detector lib (domain_tag friction clustering)
Fixes
- #347 quote heredoc delimiter so backtick in body is not command-substituted
- #339 abdication-guard counter fails toward allow-stop when unpersistable
- #338 preserve existing AGENTS.md via additive managed block (no overwrite)
Refactoring
Documentation
2026-06-26
Features
- #321 add agentic-update from-anywhere update command
- #256 cross-harness agent teams (any harness) + role-model routing
Fixes
- #337 cut per-call token cost ~30x (alwaysApply off + compat preamble + refs relocation)
- #334 clean up orphan .tmp files in stop-context atomic-write catch blocks (#262) [hooks-stack 1/3]
- #325 remove dangling ~/.local/bin/agentic-* symlinks + tighten path guards
- #322 point version-check nudge at agentic-update + /pull-and-install
- #318 --fix exits 2 when repo_dir is invalid (was silently 0)
- #290 guard identity.sh pipelines so set -e abort no longer silently kills the installer
Performance
Refactoring
Documentation
- #330 author bin/AGENTS.md and hooks/AGENTS.md (#261)
- #329 sync .agentic/config.json toggle list to authoritative source (#260)
- #324 clarify per-project opt-in/opt-out depends on the install --mode global mode
- #323 lead with updating; add agentic-update across README, updating.md, and site
- #320 fix slide overflow across all decks
- #317 document agentic-doctor health check + install safeguards
Tests
Maintenance
2026-06-25
Features
- #313 role-aware model-tier defaults + required Tier declaration field
- #310 converging symlinks in cursor + opencode install.sh
- #309 guard against second-clone split-brain
- #308 skip adapter rebuild when no adapter-source changed (consumer fast-path)
- #306 converging symlinks + guarded repo_dir write (self-heal drift)
- #305 abdication-guard Stop hook (mechanizes proactive-autonomy)
Refactoring
Documentation
- #316 replace stale agent-methodology.md references with current rules structure
- #315 align agent-team roster with full 17-agent set
- #312 pull-and-install runs agentic-doctor; document repo-dir.sh canonical resolver
- #307 codify the digest-return pattern (lean alternative to a nested-orchestration tier)
- #300 align slide decks + docs with current methodology; fix how-it-works slide-7 overflow
2026-06-24
Features
- #304 add agentic-doctor install-health check with --fix
- #303 add repo-dir.sh canonical resolver lib
- #299 revision-discipline, grounding, and verification-footer rules
- #296 in-session capture-gap nudge via PostToolUse(Task)
- #228 daemon-driven deferred /wrap — opt-in, Claude-only
Fixes
2026-06-22
Features
- #294 cross-harness agent teams (extends #256)
- #291 zero-opt-in graph signal + autonomous graphify update refresh
- #289 opt-in graphify_risk_signals - graph-derived escalation for conductor risk classification
- #288 optional Graphify graph-aware blast-radius via capability-preflight
Fixes
Documentation
2026-06-21
Uncategorized
- #287 DS-28: trigger catalog + open-goal loop contract
- #286 DS-29: operator-confirmed outcome rubric + verification_type front-of-loop artifact
- #285 DS-30: lead PR body with QA evidence for behavior-visible units + diff-read guardrail
- #284 DS-27: enforce orchestrator-singularity (deny subagent-issued Task spawns)
- #283 AUT-202: agent tool-ergonomics standard (token-efficient output, CLI over MCP)
2026-06-20
Features
- #277 freeze header + first two columns, widen data columns
- #276 gate /update-agentic-engineering to in-repo sessions with handoff
Fixes
- #279 rebuild all adapters in /update-agentic-engineering Step 3
- #278 restore DinoStack cyan tint + stack Table C status notes
Documentation
Maintenance
2026-06-19
Features
- #257 automated changelog generator + daily publish
- #249 role-model routing with cross-family antagonist reviewers
Fixes
Documentation
- #272 add local-branch lifecycle / prune to worktree management
- #270 refocus README on getting started, add logo banner
- #254 document committed-vs-local doc homes for new features
Maintenance
- #271 strip maintainer eval-harness internals from public MEMORY.md
- #252 rebuild .openclaw to sync clone-dir refs
- #251 enforce docs/planning gitignore via tracked-file guard
- #237 update remaining clone-dir refs from agentic-engineering to DinoStack
Uncategorized
2026-06-17
Features
- #235 add non-interactive install-all.sh for all adapters
- #234 collect developer identity in all 9 adapter installers (shared helper)
- #233 clearer one-finding-per-line sign-off Findings format
Fixes
Maintenance
2026-06-16
Features
- #230 add /pull-and-install (agent-guided update + fresh install)
- #229 collect developer identity in the guided installer
Fixes
Documentation
Maintenance
2026-06-14
Documentation
Maintenance
2026-06-13
Features
- #224 promote Hermes to an official tracked + CI-wired adapter (9th)
- #220 capture-gap backstop [U6]
- #219 Phase 6 capture-sweep + wrap-ticket KNW + protocol crossref [U5,U7]
- #218 dual schema + mandatory capture gate + agent KNW mapping [U2-U4]
- #214 session_uuid on emits + tool_failure_workaround event [U0]
- #213 Planning Artifacts infographic full-width + Ecosystem Comparison section
- #206 distribute 7 DinoStack infographics into topical sections
- #202 seed + commit .agentic/learnings.md via additive negation [u2+u9]
Fixes
- #211 re-track .agentic/learnings.md untracked by #203 [u3c]
- #210 restore .agentic/learnings.md carve-out clobbered by #203 [u3b]
Documentation
- #222 clarify DinoStack brand vs agentic-engineering package naming
- #221 recategorize commands + restack panels
- #217 add capture-classification.md [U1]
- #216 fix staleness across 14 decks + reorder deck index for learner flow
- #215 DinoStack-only skill inventory with operator/methodology delineation
- #212 reconcile knowledge-routing docs with shipped state [u10]
- #209 document autoMemoryDirectory intent [u8]
- #208 track .agentic/learnings.md via negation carve-out [u7]
- #207 correct learnings pipeline wiring claim [u6]
- #205 rebrand AE to DinoStack and widen for readability
Tests
Maintenance
2026-06-12
Features
- #197 commit session-log at PR Phase 8 + docs + adapters [DS-15 unit 2]
- #192 project-scope override with 4-tier resolution [DS handle override]
- #191 fix auto-harness keep gate + eval substrate (sign-flip permutation gate, parallel runner, sonnet-pinned)
- #189 add docs.dinostack.ai/install.sh vanity redirect [DS-14]
- #186 adopt Apache-2.0 (LICENSE + NOTICE) [DS-2]
Fixes
Documentation
- #199 scope .agentic/memory.md as /wrap-internal scratch [u5]
- #198 point License section at Apache-2.0 LICENSE [DS-2]
- #194 add Identity and Telemetry section [readme-docs]
- #193 document project-scope override + rebuild adapters [content-rebuild]
Maintenance
2026-06-10
Fixes
Documentation
2026-06-09
Documentation
- #180 point docs-site deploy at dinostack-docs (space-dinos-sandbox)
- #179 reskin docs site with Iridium dark look and feel
- #178 rebrand product name to DinoStack in reader-facing docs
Uncategorized
2026-06-05
Features
Documentation
Maintenance
2026-06-04
Features
- #171 docs + adapters for ae-identity-v1 (content, PR2)
- #170 auto-identity + pending buffer + operator rollup (behavioral, ae-identity-v1 PR1)
- #166 SessionStart newer-version-available notice
Fixes
Maintenance
2026-06-02
Features
Documentation
Maintenance
- #165 ignore @marp-team/* in dependabot (pinned, coordinated manually)
- #158 bump tmp to 0.2.7 and add scoped dependabot config
- #131 fix tmp CVE in scripts/
Uncategorized
2026-05-31
Documentation
Maintenance
2026-05-28
Features
- #153 motion_aware + storybook_version + blocking-mode default (P2 U4)
- #152 motion procedure + SB6 URL branching (P2 U2)
- #151 populate capabilities in 8 agent files (P2 U5)
- #150 motion enforcement + theme allowlist expansion (P2 U3)
- #149 add motion method + theme allowlist split (P2 U1)
- #147 add retro subcommand for pre-telemetry historical rollups
- #146 manifest-driven project scaffolding auto-sync
- #144 theme-aware + storybook scenario procedures (P1 U2)
- #143 reference doc + engineer addendum + Skeptic findings (P1 U3)
- #142 add theme + story_id qa_criteria fields (P1 U1)
- #140 seed FE/QA defaults + conventions polish (U4)
- #139 manifest schema + conductor preflight (U3)
- #138 accessibility, perceptual_diff, viewport procedures (U2)
- #137 add accessibility, perceptual_diff, viewport qa_criteria fields (U1)
- #136 Stage 1 committed per-developer session log
Documentation
- #148 P2 brief - motion, SB6, capability blocking-mode
- #141 ADR for project scaffolding auto-sync
- #135 telemetry install gap analysis
- #134 ADR for team-level telemetry aggregation
- #132 add External Comment Discipline rule for PR and tracker comments
Maintenance
2026-05-24
Features
Uncategorized
2026-05-23
Features
2026-05-21
Features
2026-05-20
Features
Fixes
Documentation
2026-05-19
Features
- #113 conductor never edits shippable artifacts directly (always-worktrees rule)
- #112 deterministic Marp build pipeline + slides-sync CI gate
Fixes
- #114 portable relative symlinks in install.sh
- #110 move findings.md to .agentic/ with gitignore exception
Documentation
2026-05-18
Features
- #109 curl|bash bootstrap installer + location-aware /update-agentic-engineering
- #106 enforced doc-sync obligation (docs stay aligned with code/reality)
- #105 advisory adapter-drift gate + wire .pi + close .pi prune false-negative
Fixes
Documentation
Maintenance
2026-05-16
Features
- #103 GSD-steal improvements - overview onboarding layer, debugger-on-quality-gate, .agentic/config.json
- #101 Brief-informs-Architect, pre-architect ambiguity scan, cross-artifact alignment check
- #100 add AST-grep guidance to investigator and debugger
- #98 add Kimi CLI backend support to skill-comparison harness
Fixes
Maintenance
2026-05-14
Features
- #96 add per-task dockerfile routing and py310 image
- #94 add preflight smoke check for skill-comparison harness
- #93 enforce DRY and abstraction in engineer self-check and skeptic review
- #92 add learnings-agent for session-scoped background learning capture
- #91 long-session context retention fixtures (Gap 4)
- #90 agentic-memory retrieval CLI (Gap 3)
Documentation
2026-05-13
Features
Fixes
2026-05-12
Features
- #74 add --tasks and --max-cells CLI filters for bounded smoke runs
- #72 SWE-bench fixture seeding
- #69 7 condition spec YAMLs for 8-condition eval matrix
- #68 runner + scoring for 8-condition eval matrix
- #67 frozen task corpus - 12 SWE-bench-lite tasks
- #66 add engineer component manifest and scorer
- #65 aggregator + config_discovery for 8-condition matrix
- #64 canary unit - ae_rules_payload + assert_canary + real transcript
Fixes
- #84 force docker rmi before build when --rebuild-image is set
- #83 switch Tier 3 Docker base to Python 3.9 for SWE-bench legacy compat
- #82 add requests deps to Dockerfile.swebench
- #81 inject PYTHONPATH=/workspace/repo in score-phase container
- #80 pass held_out_from_fix_dir=True so score phase mounts test files
- #79 convert unittest-style fail_to_pass IDs to pytest node-id format
- #78 forward fail_to_pass to score_cell and use git diff for engineer diffs
- #77 raise max_turns to 40, fix pytest 0-tests-collected in score phase
- #76 smoke v4 bugs - empty transcript and pytest 0 tests
- #75 add pytest-timeout, --rebuild-image flag, and transcript persistence
- #73 build Docker image once per run_matrix, not once per cell
- #71 sys.executable for pytest + score_error status on scoring failure
Documentation
2026-05-11
Features
- #63 Tier 3 Docker isolator for SWE-bench fix+score phases
- #62 scaffold evals/skill-comparison/ directory tree
Documentation
2026-05-09
Fixes
Uncategorized
2026-05-08
Uncategorized
2026-05-07
Features
Fixes
Uncategorized
2026-05-01
Features
- #52 align methodology with parallel-by-default philosophy
- #49 inject corpus-level dimension signal into editor brief
- #44 interactive planning command + downstream wiring
- #41 switch keep-metric from median-of-medians to mean-of-medians
- #39 Wave 2 adapter migrations + SKILL.md dedup
Fixes
Documentation
- #53 regenerate infographics with full label fidelity
- #43 add planning-tier slide deck + partial-rebuild deploy guidance
- #42 mirror Planning Artifacts methodology change in hub html
- #37 replace SVG infographics with rendered illustrations
Maintenance
- #51 formalize conductor-vs-subagent worktree mental model
- #48 trigger redeploy under fullmetalblanket identity (rebase)
- #47 trigger redeploy under fullmetalblanket identity
- #46 trigger redeploy with team-authorized author
- #45 rebaseline methodology SHA after PR #44
- #40 post-Wave-2 cleanup
- #38 rename agent-methodology -> METHODOLOGY across adapters and content (Wave 1.5)
2026-04-30
Breaking Changes
- #23 split agent-methodology.md into composable sections (Wave 1)
- #20 cut findings-flywheel promotion; keep regression-test obligation
- #19 remove tight-fix-path Skeptic skip
Features
- #36 canonical Confidence values (auto-harness +0.07)
- #33 planning-tier artifacts (Brief / Plan) with verification gate
- #32 first-activation notice + /agentic-status + /agentic-disable (Gap 3)
- #31 Skeptic calibration drift detection (Gap 2)
- #29 cost & latency observability (Gap 1)
- #22 add spawn presets feature
- #21 add .agentic/events.jsonl events log spec
Documentation
Maintenance
- #35 re-baseline conductor with full 20 fixtures
- #34 re-baseline debugger with full 5 fixtures
- #30 sync public docs with cost-observability changes
- #28 sync public docs and slides with tiered manifest enforcement
- #27 tiered Skeptic enforcement for module manifests
- #18 demote module-manifest enforcement to Suggestion
- #17 extract multi-dev coordination
2026-04-29
Features
Documentation
Maintenance
2026-04-28
Features
- #10 auto-inject session context and memory via @-imports
- #8 add context-mode guidance to code-standards
Fixes
- #12 preserve Claude CLI auth in Tier 2 HOME redirect
- #9 ensure rules files always load via @ imports (Claude Code) and add missing module-manifest to Codex
- #5 claude install script error
Maintenance
- #13 update MEMORY.md after Tier 2 auth fix
- #11 move compression-state.json from .claude/ to .agentic/