super-looper: v0.8.0

0.8.0 (2026-07-14)

Features

• run: add dual-host workflow core (#50) (eae6d15)

super-looper: v0.7.0

0.7.0 (2026-07-12)

Features

• plan: cap paid image calls per run with --max-images (#41) (d3af2d4)
• plan: refine a filled image slot in place with --edit (e983447)
• plan: wire reciprocal forward refs from a plan's back refs (#43) (ca6514e)

Bug Fixes

• doc-review: add an HTML-safe mutation path (#44) (2c9fbc6)
• loop: hash the plan's goal content, not its bytes (#45) (22a20fc)
• resolve-pr-feedback: stop on non-substantive rounds, not on silence (#47) (4f9b059)

super-looper: v0.6.1

0.6.1 (2026-07-12)

Bug Fixes

• resolve-pr-feedback: auto-refresh a bot-blocked review gate after quiescence (3b286ba)
• resolve-pr-feedback: auto-refresh a bot-blocked review gate after quiescence (cc97744)

super-looper: v0.6.0

0.6.0 (2026-07-11)

Features

• sl-plan: add bundled image generation script with script-side base64 injection (8d8e623)
• sl-plan: add canonical HTML plan template adapted from planf3 (3a11aae)
• sl-plan: declare HTML plans stateful tracked artifacts in section contract (198116a)
• sl-plan: planf3-style HTML plans with images and live tracking (dd193e8)
• sl-plan: stamp HTML plans from canonical template and fill image slots post-write (b4efb65)
• sl-work: maintain advisory status markers in HTML plans during interactive execution (1a91d74)

Bug Fixes

• sl-plan: harden image-slot grammar, script failure paths, and fill-later routing per review (3649eff)
• sl-work: release finished subagents so they mark complete in the UI (fd9b4de)
• sl-work: scope ship-time metadata ownership and own global validation markers (baec94f)

super-looper: v0.5.0

0.5.0 (2026-07-05)

Features

• evals: behavioral eval suites for the four load-bearing skills (61393d7)
• goal guard, loop resume, learning gate, hygiene, and evals (2bcd93d)
• hooks: plugin goal-guard hook — deny goal-file writes in unattended runs (7f03775)
• learn: evaluator gate separates generator from evaluator on the learning path (2b643f2)
• learn: refresh-due threshold signal and corpus staleness pass (dc45641)
• lfg: run-progress file written at every step boundary (29d7da2)
• lfg: write the goal-change protocol as a grep-tested contract (0897804)
• loop: instrument goal_fidelity from review verdict to run-record ledger (6ae382b)
• loop: resume retries from the recorded step instead of cold-restarting (5d2038f)
• solutions: add confidence, provenance, and evidence trust fields (665300c)

Bug Fixes

• agents: retire two unwired agents and enforce fleet color conventions (feb08d3)
• docs: generate docs/skills index from frontmatter with integrity tests (ba112b7)
• release: clean stale packaging artifacts and close version-drift validator gap (ab9795f)
• review: apply review findings (61ce166)

super-looper: v0.4.1

0.4.1 (2026-06-26)

Bug Fixes

• agents: assign role-based colors to all 43 agents (c43ffc1)
• agents: role-based, fleet-scoped color scheme for all 43 agents (a031cfc)

super-looper: v0.4.0

0.4.0 (2026-06-21)

Features

• pulse: wire unattended completion rate to a run-record ledger (db4dd7d)
• sl-product-pulse: add local JSONL ledger source kind (U3) (23cd9d5)
• sl-product-pulse: render learning_reuse as labeled git citation proxy (U5) (61eaf1b)

Bug Fixes

• loop: run ledger append on failed loops and guard ledger integrity (f261e38)

super-looper: v0.3.0

0.3.0 (2026-06-20)

Features

• capture learnings at the autopilot's ship-time seam (8cc067f)
• clean plan-to-work handoff for the implementation autopilot (3460218)
• lfg: add debug-escalation rung to the CI-fix loop (8746b77)
• lfg: add debug-escalation rung to the CI-fix loop (3e4ef45)
• lfg: trigger the sl-learn seam after CI green, before DONE (29e4ba7)
• sl-learn: add ship-time learning-capture seam (1b15a81)
• sl-plan: offer the work loop at the end-of-plan seam (855ecb4)

Bug Fixes

• lfg: accept HTML plan shape at the plan-input gate (#11) (0e3b035)
• README: restore sl-handoff docs link (#12) (af3c754)
• resolve-pr-feedback: add wait-for-bot-review poll script (30f4f1c)
• resolve-pr-feedback: harden quiescence gate against gh failure modes (fefb41f)
• resolve-pr-feedback: make reply-to-pr-thread example a single pinned command (dafb4ed)
• resolve-pr-feedback: surface wait-for-bot-review and scope it out of targeted mode (614daa1)
• resolve-pr-feedback: wait for async bot re-review before concluding verify loop (02db673)
• resolve-pr-feedback: wait for bot re-review before concluding verify loop (d96850f)
• review: disambiguate plan: marker from handoff context, correct plan-input docs, add coverage (7051dcd)

super-looper: v0.2.0

0.2.0 (2026-06-17)

Features

• solutions: single-source schema for docs/solutions frontmatter (9029e76)

Bug Fixes

• solutions: address PR review feedback (#2) (c9c7e8b)
• solutions: regenerate schema docs from web-stack enums and lock with drift gate (0035a8c)