Open Issues Need Help
View All on GitHub ci: ruff 0.17.3 left three pins at 0.16.10, so main fails its lockstep tests and pre-commit disagrees with CI about 3 hours ago
good first issue ci dependencies hacktoberfest
documentation good first issue hacktoberfest
tests: cover _resolve_via_ledger — the whole ledger-anchored idempotency derivation is unexercised about 24 hours ago
good first issue tests hacktoberfest
good first issue tests hacktoberfest
tests: cover the thin adapters' availability probes, non-dict arguments, and unknown-token no-ops 1 day ago
good first issue adapter tests hacktoberfest
tests: drive the TUI detail view's keyboard control surface (tabs, selection, queue mutations) 1 day ago
good first issue tests hacktoberfest
documentation good first issue hacktoberfest
good first issue tests hacktoberfest
tests: BrowserAdapter.navigate body is uncovered and the only test skips when playwright is present 8 days ago
good first issue adapter tests
good first issue tests
bug good first issue cli tests
good first issue tests
bug good first issue tests
Add CONTINUUM to awesome-ai-plugins? 11 days ago
documentation good first issue
good first issue adapter state tests
tests: cover _parse_authority_verdict and built-in probe config validation in reconcilers.py 11 days ago
good first issue tests
good first issue tests
good first issue state tests
mcp doctor: an observed client name with spaces is recorded as several names the host never declared 11 days ago
bug good first issue mcp
benchmark: phase6 reports stamp naive local time, while every other timestamp in the tree is UTC-aware 11 days ago
bug good first issue benchmark
good first issue tests
good first issue tests
good first issue tests
good first issue tests
tests: cover _approval_earlier and the conformance branches in plugins/validation_rules.py 16 days ago
good first issue tests
good first issue tests
good first issue tests
tests: add unit tests for the OpenAI adapter's _format_annotation and context extraction 17 days ago
good first issue tests
tests: add SDK-free unit tests for the LangChain adapter helpers (_extract_run_id and friends) 17 days ago
good first issue tests
tests: add unit coverage for the serve sidecar wire client (SidecarClient, run_serve, cmd_serve) 17 days ago
good first issue tests
documentation good first issue hacktoberfest
docs: the translated READMEs are missing the Documentation map section and two Quick links entries 17 days ago
documentation good first issue hacktoberfest
good first issue tests
good first issue hacktoberfest
documentation good first issue
documentation good first issue tests
good first issue tests ci
good first issue tests
documentation good first issue
documentation good first issue
documentation good first issue
documentation good first issue
documentation good first issue
documentation good first issue
documentation good first issue
documentation good first issue
documentation good first issue ci
docs: references/install.md cites seven pyproject.toml lines that do not point at the extras 20 days ago
documentation good first issue
tests: resolve_compaction_bound is the least covered module in the package and guards the anchor 20 days ago
good first issue tests hacktoberfest
tests: KubernetesAdapter.run_job body is uncovered and available() has four untested branches 22 days ago
good first issue tests hacktoberfest
packaging: playwright and kubernetes are promised in mypy overrides but declared as no extra 22 days ago
good first issue hacktoberfest
good first issue
tests: ContainerAdapter.run_in_container body is uncovered and the only test skips when docker is present 22 days ago
good first issue tests hacktoberfest
tests: silence_seconds has zero coverage and the cadence contract validator never raises 22 days ago
good first issue tests hacktoberfest
good first issue tests hacktoberfest
good first issue tests hacktoberfest
documentation good first issue hacktoberfest
documentation good first issue hacktoberfest
documentation good first issue hacktoberfest
documentation good first issue hacktoberfest
documentation good first issue hacktoberfest
documentation good first issue hacktoberfest
five more modules omit public names from __all__: reconciliation, snapshot, policy, registry, models 23 days ago
enhancement good first issue hacktoberfest
good first issue tests
docs: adapters.md documents an engine= keyword that the LangGraph, OpenAI and LangChain adapters reject about 1 month ago
documentation good first issue adapter
after #1064 (benchmarks/), scripts/ and demo-run/ are the last Python directories outside every lint gate, carrying 9 findings and formatting drift about 1 month ago
good first issue ci
enhancement good first issue hacktoberfest
documentation good first issue hacktoberfest
docs and references claim 29 event types; the EventType enum has 51 (seven sites, two asserting a complete list) about 1 month ago
documentation good first issue hacktoberfest
pin-marker surface contradicts itself: docs name pin_markers_for_state as the emitter, but context.py imports the private _pin_marker and the helper has no caller about 1 month ago
documentation enhancement good first issue hacktoberfest
Follow-up to #392: recovery/derived.py is never imported, so both derived-artifact producers inline the stamping and the invariant check has no consumer about 1 month ago
enhancement good first issue provenance
CheckpointManager.checkpoint_on_recovery is never called: RECOVERY anchors and the last_recovery_anchor lookup are dead surface about 1 month ago
enhancement good first issue
ActionLedger.compensate has no transport: ACTION_COMPENSATED is unreachable and the compensations counter can never be non-zero about 1 month ago
enhancement good first issue
restore_to_anchor and merge_to_anchor are exported with zero callers, and restore_to_anchor silently discards its reason parameter about 1 month ago
enhancement good first issue
benchmarks/ is imported by tests but excluded from every lint gate, carrying 9 findings about 1 month ago
good first issue ci
docs: README and references/cli.md say every command accepts --json, but the trailing position is rejected with exit 2 about 2 months ago
documentation good first issue hacktoberfest
references/testing.md and install.md cite stale test counts (~2,160 and ~1,300) that sit outside the docs-count guard's three-file scope about 2 months ago
documentation good first issue hacktoberfest
docs: README's Origin section says 1380 tests while the same file's guarded figure is 2,241, and the count guard cannot see the claim about 2 months ago
documentation good first issue tests
24 modules ship public names but no __all__, so the repo's own export convention does not reach them about 2 months ago
enhancement good first issue
tests: is_quiet_window and render_trajectory_report are the trajectory report's trigger and renderer, and neither has a direct test about 2 months ago
good first issue tests
docs: human_steps is a computed top-level key, not an attribute of RecoveryDecision or RecoveryContract, and five doc sites teach the broken access about 2 months ago
documentation good first issue
tests: the idempotency token helpers that carry the #365 double-effect fix have no direct test about 2 months ago
good first issue tests
ci: _bugaudit/ is the last tracked Python directory outside every lint gate, with 2 findings and formatting drift about 2 months ago
good first issue ci
docs: CONTINUUM_MCP_SLIM silently deletes 9 of the 12 MCP tools and no document mentions it about 2 months ago
documentation good first issue mcp
docs: four code pointers in signing_key_rotation, threat_model and multi_agent_isolation land on blank or unrelated lines about 2 months ago
documentation good first issue
dashboard: pending_actions_with_keys omits REQUIRES_REVIEW, so escalated actions cannot be settled from the UI about 2 months ago
bug good first issue
CITATION.cff pins version 0.1.0 while the package is 0.1.2, and CONTRIBUTING says the version lives in two places when the #838 guard checks four about 2 months ago
documentation good first issue hacktoberfest
backoff_delay is exported in budgets.__all__ but nothing ever paces a retry with it about 2 months ago
enhancement good first issue
the build-your-own-dashboard guide reads decision.human_steps, which does not exist on RecoveryDecision: every copy-paste example raises AttributeError about 2 months ago
bug documentation good first issue
references/cli.md documents the dashboard default port as 8080; the CLI and the dashboard default to 8000 about 2 months ago
documentation good first issue hacktoberfest
DuplicateAction and LeaseError are exported exceptions that no code path can raise about 2 months ago
enhancement good first issue
RewindError and LedgerLockError error paths have no direct test: the rewind suite is happy-path only and the ledger lock-contention raise is never asserted about 2 months ago
good first issue tests hacktoberfest
load_reconcilers silently returns {} for a registry with no probes key about 2 months ago
enhancement good first issue
replay_similarity backends are built and tested but never wired into the replay guard about 2 months ago
enhancement help wanted tests
bench(horizon): abort_condition_year cannot pass by construction, deflating published accuracy to 0.60 about 2 months ago
bug help wanted benchmark
observability metrics API is dead: get_metrics/reset_metrics/collect_from_decision have no caller about 2 months ago
enhancement good first issue tests
Homebrew tap for the continuum CLI (brew install path alongside uvx/pipx) about 2 months ago
help wanted
recovery verdict is advisory by default: RecoveryDecision.permits has no product caller about 2 months ago
documentation enhancement help wanted
security/trust_gate.py and security/revalidation.py (Extensions 1 and 2) are built but never wired into the engine about 2 months ago
enhancement help wanted research security