Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
51 commits
Select commit Hold shift + click to select a range
8d4cce8
docs: review evidence reliability and plan bounded P0 work
k2jac9 Nov 15, 2026
9141d8f
build: freeze the tested dependency resolution
k2jac9 Nov 15, 2026
0d7f61c
feat: explain current eligibility through the shared read gate
k2jac9 Nov 15, 2026
c322e57
docs: define autonomous reliability milestones and decision boundaries
k2jac9 Nov 15, 2026
cfa8d3c
fix: reject malformed accepted output in eligibility gate
k2jac9 Nov 15, 2026
1509c38
feat: atomically revise admissions at an expected epoch
k2jac9 Nov 15, 2026
261a863
feat: add pure versioned incident protocol and exact-byte digests
k2jac9 Nov 15, 2026
ef42bcf
fix: bound incident protocol allocation before size refusal
k2jac9 Nov 15, 2026
94a6d43
feat: exercise incident preparation through Store timed gates
k2jac9 Nov 15, 2026
0dd48f8
test: verify Store incident byte parity in CI
k2jac9 Nov 15, 2026
3b45721
docs: bind replay safeguards and SQLite lifetime repair
k2jac9 Nov 15, 2026
062725f
fix: close SQLite audit and controller connections
k2jac9 Nov 15, 2026
8a519e1
docs: define deterministic decision bundle digest encoding
k2jac9 Nov 15, 2026
43556ae
feat: define bounded pure decision bundle replay
k2jac9 Nov 15, 2026
b5756e0
fix: enforce recovery bundle ledger invariants
k2jac9 Nov 15, 2026
37f25c4
feat: authorize decision exports within one Store snapshot
k2jac9 Nov 15, 2026
b921743
test: exercise decision replay across process boundaries
k2jac9 Nov 15, 2026
04cb3b6
feat: validate complete comparable policy evaluation records
k2jac9 Nov 15, 2026
62f5646
fix: validate partial policy facts and legacy attempt timing
k2jac9 Nov 15, 2026
0923e13
docs: bind policy migration and activation integration scopes
k2jac9 Nov 15, 2026
2cc5f9d
feat: persist immutable scoped policy versions and evaluations
k2jac9 Nov 15, 2026
f1a0af0
fix: prevent rowid replacement of immutable policy records
k2jac9 Nov 15, 2026
33b3303
docs: bind activation consistency and diagnostic runtime boundaries
k2jac9 Nov 16, 2026
ee5beb6
feat: atomically activate evaluated policies
k2jac9 Nov 16, 2026
a8f8c4b
test: demonstrate policy activation and rollback without model calls
k2jac9 Nov 16, 2026
9329950
fix: bound policy fixture input before deserialization
k2jac9 Nov 16, 2026
a621ad8
docs: retain policy activation fixture packet
k2jac9 Nov 16, 2026
31e0dd1
feat: plan and score bounded offline ablations
k2jac9 Nov 16, 2026
794f64e
fix: preserve unspecified-unit accounting observations
k2jac9 Nov 16, 2026
5bf3c75
Add source-bound isolated runtime matrix verifier and refusal tests
k2jac9 Nov 16, 2026
26045ce
Bind runtime receipts to the complete qualification manifest
k2jac9 Nov 16, 2026
353ff2f
Retain empty log hashes alongside nonempty qualification results
k2jac9 Nov 16, 2026
ab8e162
Normalize runtime receipt counts across JSON persistence
k2jac9 Nov 16, 2026
dd52954
Bind persisted runtime command and child interpreter provenance
k2jac9 Nov 16, 2026
fce095d
docs: record isolated runtime qualification and rollback evidence
k2jac9 Nov 16, 2026
d60159d
fix: isolate unavailable runtime lanes at the matrix CLI
k2jac9 Nov 16, 2026
7684b4f
docs: record isolated lane fix and five-runtime requalification
k2jac9 Nov 16, 2026
ae4aa08
Pin qualified project runtimes and add verified local launcher
k2jac9 Nov 16, 2026
64e0b1b
Bind native helper resolution and reject Cargo selector bypasses
k2jac9 Nov 16, 2026
9d94dbf
Record final runtime launcher qualification and activation handoff
k2jac9 Nov 16, 2026
e32c968
Reject inherited native flag families and target-specific selectors
k2jac9 Nov 16, 2026
1870b60
Record native boundary correction verification and retained warnings
k2jac9 Nov 16, 2026
d370b41
Clarify units of the retained native wrapper audit
k2jac9 Nov 16, 2026
5e9c76e
Bind rustdoc selection and version unknown-unit replay observations
k2jac9 Nov 16, 2026
0899eca
Document V2 selection and replay semantics with final source evidence
k2jac9 Nov 16, 2026
7c7ca36
docs: remove unrelated integration references
k2jac9 Nov 16, 2026
6278767
Separate prepared adapter operations and preserve existing verificati…
k2jac9 Nov 16, 2026
c7501c4
Document warning cleanup and log preservation verification
k2jac9 Nov 16, 2026
e43677d
docs: map reliability commits and verification boundaries for review
k2jac9 Nov 16, 2026
c7703e1
fix: isolate hosted test environments and pin current checkout action
k2jac9 Nov 16, 2026
afd8c5d
Make prepared worker hang test wait for a durable claim
k2jac9 Nov 16, 2026
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
142 changes: 127 additions & 15 deletions .github/workflows/check.yml
Original file line number Diff line number Diff line change
Expand Up @@ -6,53 +6,150 @@ permissions:

jobs:
contracts:
name: contracts (${{ matrix.role }})
runs-on: ubuntu-latest
strategy:
fail-fast: false
matrix:
include:
- role: stable
rust: '1.100.0'
rust-release: '1.100.0'
rust-commit: '0ef1bc386059596dada8fb65b50425aab2ea637a'
cargo-version: 'cargo 1.100.0 (3acc79c0c 2026-11-05)'
python: '3.15.2'
python-version: '3.15.2'
- role: rust-upcoming
rust: 'nightly-2026-11-16'
rust-release: '1.102.0-nightly'
rust-commit: '5f4effd579500175af844241eb179e756190f1e7'
cargo-version: 'cargo 1.102.0-nightly (ede6cefea 2026-11-11)'
python: '3.15.2'
python-version: '3.15.2'
# This pair also retains the accepted local development baseline.
- role: rust-previous-and-retained-baseline
rust: '1.99.1'
rust-release: '1.99.1'
rust-commit: '58cf8f9c35b355199727627d07ae34a4e2846d67'
cargo-version: 'cargo 1.99.1 (5f94df478 2026-08-27)'
python: '3.15.2'
python-version: '3.15.2'
- role: python-upcoming
rust: '1.100.0'
rust-release: '1.100.0'
rust-commit: '0ef1bc386059596dada8fb65b50425aab2ea637a'
cargo-version: 'cargo 1.100.0 (3acc79c0c 2026-11-05)'
python: '3.16.0-alpha.2'
python-version: '3.16.0a2'
- role: python-previous
rust: '1.100.0'
rust-release: '1.100.0'
rust-commit: '0ef1bc386059596dada8fb65b50425aab2ea637a'
cargo-version: 'cargo 1.100.0 (3acc79c0c 2026-11-05)'
python: '3.14.10'
python-version: '3.14.10'
timeout-minutes: 10
env:
KAIROTENSOR_REQUIRE_LINUX_CLOCK: '1'
# Explicitly overrides rust-toolchain.toml in every compatibility row.
RUSTUP_TOOLCHAIN: ${{ matrix.rust }}
RUSTUP_AUTO_INSTALL: '0'
EXPECTED_RUST_RELEASE: ${{ matrix.rust-release }}
EXPECTED_RUST_COMMIT: ${{ matrix.rust-commit }}
EXPECTED_CARGO_VERSION: ${{ matrix.cargo-version }}
EXPECTED_PYTHON_VERSION: ${{ matrix.python-version }}
steps:
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
- uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
with:
persist-credentials: false
- name: Select exact Rust lane
uses: dtolnay/rust-toolchain@df2a88e9fe4098bad09cb42c3dec7187f15861fc
with:
toolchain: ${{ matrix.rust }}
- name: Select exact Python lane
uses: actions/setup-python@6d4fd22b20465ccc0f7354476588e8df7c9ca0e0 # v7.1.0
with:
python-version: ${{ matrix.python }}
check-latest: false
- name: Verify actual Rust, Cargo and parent/child Python identities
shell: bash
run: |
set -euo pipefail
rustc -vV
cargo --version
rustup which rustc
rustup which cargo
python3 - <<'PYTHON'
import json, os, pathlib, subprocess, sys
rust = dict(line.split(': ', 1) for line in subprocess.check_output(['rustc', '-vV'], text=True).splitlines()[1:])
assert rust['release'] == os.environ['EXPECTED_RUST_RELEASE'], rust
assert rust['commit-hash'] == os.environ['EXPECTED_RUST_COMMIT'], rust
assert subprocess.check_output(['cargo', '--version'], text=True).strip() == os.environ['EXPECTED_CARGO_VERSION']
assert sys.version.split()[0] == os.environ['EXPECTED_PYTHON_VERSION'], sys.version
child = json.loads(subprocess.check_output(['python3', '-c', 'import json,sys,os;print(json.dumps([os.path.realpath(sys.executable),sys.version.split()[0]]))'], text=True))
assert child == [str(pathlib.Path(sys.executable).resolve()), os.environ['EXPECTED_PYTHON_VERSION']], child
print(json.dumps({'parent': str(pathlib.Path(sys.executable).resolve()), 'child': child}))
PYTHON
- name: Test the Rust store, including the withdrawal example
run: cargo test --all-targets
run: cargo test --locked --all-targets
- name: Run the revision and model handoff examples
run: |
cargo run --example withdrawal
cargo run --example deterministic_verdict
cargo run --example deadline_aware_recovery
cargo run --example runtime_recovery
cargo run --example selective_revision
cargo run --example corrected_handoff
cargo run --example scoped_model_reuse
cargo run --example scoped_model_reuse -- --replay-invalid
cargo run --example token_cap_eval
cargo run --example budgeted_recovery
cargo run --locked --example withdrawal
cargo run --locked --example deterministic_verdict
cargo run --locked --example deadline_aware_recovery
cargo run --locked --example runtime_recovery
cargo run --locked --example selective_revision
cargo run --locked --example corrected_handoff
cargo run --locked --example scoped_model_reuse
cargo run --locked --example scoped_model_reuse -- --replay-invalid
cargo run --locked --example token_cap_eval
cargo run --locked --example budgeted_recovery
- name: Check the original reference contracts
run: |
python3 -m unittest discover -s reference -p test_memory.py
python3 reference/belief_oracle.py
python3 -m unittest discover -s scripts -p 'test_*.py'
cargo build --locked --offline --example recovery_policy_eval --example decision_replay
KAIROTENSOR_RECOVERY_RUNNER="$PWD/target/debug/examples/recovery_policy_eval" KAIROTENSOR_DECISION_REPLAY="$PWD/target/debug/examples/decision_replay" python3 -m unittest discover -s scripts -p 'test_*.py'
- name: Verify bounded offline ablation planning and recorded scoring
run: KAIROTENSOR_DECISION_REPLAY="$PWD/target/debug/examples/decision_replay" python3 -m unittest discover -s scripts -p test_decision_ablation.py
- name: Export and replay a decision in separate offline processes
run: |
set -euo pipefail
cargo run --locked --offline --example decision_replay -- export-fixture "$RUNNER_TEMP/decision-v1.json"
cargo run --locked --offline --example decision_replay -- replay "$RUNNER_TEMP/decision-v1.json"
python3 scripts/audit_recovery_bundle.py --decision-v1 "$RUNNER_TEMP/decision-v1.json"

- name: Verify archived output-cap measurements
run: python3 scripts/summarize_token_caps.py docs/evaluations/token-caps-*-2026-11-13.json

- name: Verify synthetic recovery policy comparison
run: python3 scripts/recovery_policy_eval.py --fixture
- name: Exercise typed policy activation and rollback without model calls
run: |
python3 scripts/recovery_policy_eval.py --activation-fixture > "$RUNNER_TEMP/policy-activation-fixture.json"
cargo run --locked --offline --example policy_activation -- --fixture "$RUNNER_TEMP/policy-activation-fixture.json" --out "$RUNNER_TEMP/policy-activation"
python3 scripts/recovery_policy_eval.py --archived-v1-refusal > "$RUNNER_TEMP/archived-v1-promotion-refusal.json"
- name: Collect and score Store-backed synthetic recovery trajectories
run: |
cargo run --example recovery_policy_eval -- --trace "$RUNNER_TEMP/recovery-policy-trace.json"
cargo run --locked --example recovery_policy_eval -- --trace "$RUNNER_TEMP/recovery-policy-trace.json"
python3 scripts/recovery_policy_eval.py "$RUNNER_TEMP/recovery-policy-trace.json"
KAIROTENSOR_RECOVERY_RUNNER="$PWD/target/debug/examples/recovery_policy_eval" python3 -m unittest discover -s scripts -p test_recovery_policy_runner.py
KAIROTENSOR_RECOVERY_RUNNER="$PWD/target/debug/examples/recovery_policy_eval" python3 -m unittest discover -s scripts -p test_audit_recovery_bundle.py
- name: Exercise bounded calibration without model inference
run: python3 scripts/recovery_calibration.py --output "$RUNNER_TEMP/recovery-calibration.json"

- name: Exercise offline timed runtime through the fixture Adapter
run: cargo run --example timed_recovery_runner -- --out "$RUNNER_TEMP/timed-recovery-runner"
run: cargo run --locked --example timed_recovery_runner -- --out "$RUNNER_TEMP/timed-recovery-runner"

- name: Prepare frozen semantic calibration requests without inference
run: python3 scripts/incident_routing.py prepare --out "$RUNNER_TEMP/incident-routing-calibration"

- name: Verify all frozen incident request bytes through Store capture without inference
run: |
cargo run --locked --example incident_store_bridge -- --requests "$RUNNER_TEMP/incident-store-requests.json"
python3 scripts/check_incident_store_bridge.py --requests "$RUNNER_TEMP/incident-store-requests.json"
python3 -m unittest discover -s scripts -p test_incident_store_bridge.py

- name: Emit bounded prepared-request calibration proposal without inference
run: python3 scripts/incident_preflight.py --out "$RUNNER_TEMP/incident-preflight-plan.json"

Expand All @@ -61,3 +158,18 @@ jobs:

- name: Exercise prepared backend with fake dependencies only
run: python3 scripts/incident_transformers_controller.py fixture --out "$RUNNER_TEMP/incident-transformers-fixture"

checkout-previous:
name: checkout previous stable (v6.1.0)
runs-on: ubuntu-latest
timeout-minutes: 5
steps:
- uses: actions/checkout@d23441a48e516b6c34aea4fa41551a30e30af803 # v6.1.0
with:
persist-credentials: false
- name: Verify event checkout and required source files
shell: bash
run: |
test "$(git rev-parse HEAD)" = "$GITHUB_SHA"
test -f Cargo.lock
test -f src/lib.rs
2 changes: 1 addition & 1 deletion .github/workflows/recovery-policy.yml
Original file line number Diff line number Diff line change
Expand Up @@ -31,7 +31,7 @@ jobs:
TOKENIZERS_PARALLELISM: 'false'
OMP_NUM_THREADS: '2'
steps:
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
- uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
with:
persist-credentials: false
- uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065 # v5.6.0
Expand Down
1 change: 1 addition & 0 deletions .python-version
Original file line number Diff line number Diff line change
@@ -0,0 +1 @@
3.15.2
Loading
Loading