61241c3a82
Implement self-contained runtime: install runtime tree, git-root memory, canonical agent paths, + 4 suites
...
- install.sh: install runtime tree under OPENCODE_DEV_AGENT_TEAM
(bin/, 12 skills, improvements/, install-manifest.json), idempotent
shell-rc export, --uninstall (preserves improvements/ user data) and
--migrate; preserves 13-agent copy, .backup retention, KEEP_BACKUPS=5,
count gate, cmp -s integrity, install-time permission gate
- memory-lifecycle.sh: resolve MEMORY_DIR from project git root + OPENCODE_MEMORY_DIR
- all 13 agents: canonical runtime sentence + env-resolved runtime paths;
project-scoped refs unchanged
- tests: new test-install (15), test-runtime (10), test-path-resolution (12),
test-memory-isolation (12); TEAM_ROOT override on existing 4 suites;
test-all.sh registers 8 suites (100 checks total)
2026-09-08 10:17:38 -04:00
03be6211ed
Add test-all.sh runner; fix bootstrap idempotency flake and test isolation
...
- test-all.sh: single entry point running all four test suites (51 checks),
non-zero exit on any failure, --verbose shows failing suite output.
- repo-bootstrap.sh: write_meta now compares stable fields excluding the
second-resolution generated_at timestamp, so two runs straddling a second
boundary no longer rewrite .bootstrap-meta (T02 idempotency flake fixed).
- test-integration.sh: pre-clean memory residue before I01/I02 so checks are
independent of prior interrupted-run state.
- README.md: document test-all.sh as the top-level test entry point.
2026-09-08 06:04:47 -04:00
4465cb24b7
Phases 5-6, 10, 12, 14: task lifecycle, integration tests, proposal wiring
...
- orchestrator.md: added 'Selection with memory and skills' and
'Skip-when efficiency guards' to Dynamic Agent Selection; added
memory-recall/skill-load fields to Decomposition work items; added
full 'Task Lifecycle' stage loop (RECALL→UNDERSTAND→ESTIMATE→LOAD→
PLAN→DISPATCH→VERIFY→LEARN→STORE→IMPROVE→REPORT) with memory stage
gate; rewrote Improvement Proposals with evidence-driven detection
triggers (recurring failure, missing skill, routing inefficiency,
docs gap, process friction, new pattern), detection flow, and
concrete proposal format.
- docs/AGENT_ARCHITECTURE.md: documented detection triggers and the
task lifecycle in section 14.
- scripts/test-integration.sh: new integration suite (12 tests) —
memory-lifecycle store/recall/list/search end-to-end, session
create/list/cleanup, skill frontmatter/owner validation (incl.
multi-owner resolution), agent↔skill mapping integrity, action
catalog memory/skills integration, task lifecycle presence,
skip-when guards, improvement proposal pending→applied lifecycle,
proposal format consistency, and real-skill owner validation.
Also documents/works around grep -q SIGPIPE under pipefail.
All 51 structural/integration tests across four suites pass.
2026-09-08 05:42:26 -04:00
Your Name