COMMITS
March 26, 2026
A
chore: deprecate in favor of gentle-ai
Alan Buscaglia committed
A
fix(sync): restore gentle-ai parity across assets and installers
Alan Buscaglia committed
A
fix(orchestrator): context-inflation delegation + align skill resolution
Alan Buscaglia committed
A
perf(skills): compact shared prompts ~17% + add judgment-day skill (parity v1.8.9)
Alan Buscaglia committed
March 25, 2026
A
refactor(sdd): token optimization + executor boundary (parity with gentle-ai v1.8.3)
Alan Buscaglia committed
A
feat(opencode): self-sufficient sub-agents for skill discovery
Alan Buscaglia committed
A
fix(opencode): pass agent model through delegate() via session.prompt()
Alan Buscaglia committed
A
fix(opencode): promote @agent-name mentions to recommended model routing
Alan Buscaglia committed
March 24, 2026
A
fix(opencode): add per-agent model field and document model routing workaround
Alan Buscaglia committed
March 16, 2026
A
docs(readme): split 945-line README into lean 145-line overview + 7 docs
Alan Buscaglia committed
A
docs(orchestrator): add delegate-first rule — prefer async over sync delegation
Alan Buscaglia committed
A
feat(opencode): add background-agents plugin and unify single/multi configs
Alan Buscaglia committed
March 15, 2026
A
fix(skills): resolve final review findings — registry ownership, envelope schema, stale refs
Alan Buscaglia committed
A
fix(skills): add sdd-phase-common.md references to verify and archive skills
Alan Buscaglia committed
A
A
perf(skills): reduce token overhead ~52,400T per SDD pipeline
Alan Buscaglia committed
A
docs(orchestrator): add Hard Stop Rule to enforce delegation on all file operations
Alan Buscaglia committed
A
docs(standards): add branch naming and commit message regex validation rules
Alan Buscaglia committed
A
feat(skills): add issue-creation skill, branch-pr skill, and AGENTS.md index
Alan Buscaglia committed
A
feat(github): enforce issue-first PR workflow with labels, templates, and CI checks
Alan Buscaglia committed
March 14, 2026
A
feat(orchestrator): add skill loading to non-SDD delegation for all agents
Alan Buscaglia committed
March 13, 2026
A
docs(readme): use real provider/model-id format in multi-model example
Alan Buscaglia committed
A
fix(scripts): make OpenCode JSON merge idempotent with model preservation
Alan Buscaglia committed
A
fix(skills): restructure engram reads to grouped search-then-retrieve pattern
Alan Buscaglia committed
A
feat(opencode): add multi-model mode with one agent per SDD phase
Alan Buscaglia committed
A
fix(opencode): replace bracket placeholders with OpenCode shell syntax
Alan Buscaglia committed
A
fix(scripts): add .gitattributes to enforce LF line endings on shell scripts
Alan Buscaglia committed
March 12, 2026
A
fix(skills): add inline mode guards to prevent openspec dir creation in engram/none mode
Alan Buscaglia committed
March 11, 2026
A
fix(setup): clean up stale "agents" plural key from opencode.json
Alan Buscaglia committed
A
fix(setup): strip preamble from examples + fix OpenCode agent key
Alan Buscaglia committed