associate 0.9.1__tar.gz → 0.11.0__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- associate-0.11.0/.devague/current +1 -0
- associate-0.11.0/.devague/current_plan +1 -0
- associate-0.11.0/.devague/deliveries/associate-multi-model-on-pi-1-1-0.json +285 -0
- associate-0.11.0/.devague/frames/associate-multi-model-on-pi-1-1-0.json +1728 -0
- associate-0.11.0/.devague/plans/associate-multi-model-on-pi-1-1-0.json +1113 -0
- {associate-0.9.1 → associate-0.11.0}/.github/workflows/publish.yml +45 -0
- {associate-0.9.1 → associate-0.11.0}/.github/workflows/tests.yml +58 -7
- {associate-0.9.1 → associate-0.11.0}/.gitignore +8 -0
- {associate-0.9.1 → associate-0.11.0}/.markdownlint-cli2.yaml +6 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/index.ts +166 -42
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/lib/context.ts +54 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/lib/contract.ts +85 -1
- associate-0.11.0/.pi/extensions/associate/lib/diff.ts +275 -0
- associate-0.11.0/.pi/extensions/associate/lib/gate.ts +1247 -0
- associate-0.11.0/.pi/extensions/associate/lib/guard.ts +662 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/lib/paths.ts +10 -1
- associate-0.11.0/.pi/extensions/associate/lib/prompt.ts +270 -0
- associate-0.11.0/.pi/extensions/associate/lib/provider.ts +267 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/lib/runtime.ts +28 -1
- associate-0.11.0/.pi/extensions/associate/lib/toolset.ts +239 -0
- associate-0.11.0/.pi/extensions/associate/lib/trust.ts +25 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/lib/walk.ts +153 -8
- associate-0.11.0/.pi/extensions/associate/prompts/cortex.md +59 -0
- associate-0.11.0/.pi/extensions/associate/prompts/senses.md +57 -0
- associate-0.11.0/.pi/extensions/associate/prompts/super-vl.md +63 -0
- associate-0.11.0/.pi/extensions/associate/prompts/worker.md +47 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tests/contract.test.ts +26 -0
- associate-0.11.0/.pi/extensions/associate/tests/delegate.test.ts +922 -0
- associate-0.11.0/.pi/extensions/associate/tests/diff.test.ts +149 -0
- associate-0.11.0/.pi/extensions/associate/tests/eidetic.test.ts +290 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tests/extension.test.ts +107 -9
- associate-0.11.0/.pi/extensions/associate/tests/gate.test.ts +1324 -0
- associate-0.11.0/.pi/extensions/associate/tests/guard.test.ts +408 -0
- associate-0.11.0/.pi/extensions/associate/tests/ledger.test.ts +292 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tests/load-extension.ts +13 -5
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tests/pi-integration.test.ts +27 -14
- associate-0.11.0/.pi/extensions/associate/tests/prompt.test.ts +468 -0
- associate-0.11.0/.pi/extensions/associate/tests/provider.test.ts +432 -0
- associate-0.11.0/.pi/extensions/associate/tests/read-image.test.ts +217 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tests/search.test.ts +21 -2
- associate-0.11.0/.pi/extensions/associate/tests/spawn.test.ts +1075 -0
- associate-0.11.0/.pi/extensions/associate/tests/tests-tool.test.ts +614 -0
- associate-0.11.0/.pi/extensions/associate/tests/toolcall.test.ts +1097 -0
- associate-0.11.0/.pi/extensions/associate/tests/toolset.test.ts +227 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tests/walk.test.ts +106 -0
- associate-0.11.0/.pi/extensions/associate/tools/delegate.ts +1002 -0
- associate-0.11.0/.pi/extensions/associate/tools/eidetic.ts +388 -0
- associate-0.11.0/.pi/extensions/associate/tools/ledger.ts +348 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tools/read.ts +100 -5
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tools/search.ts +20 -6
- associate-0.11.0/.pi/extensions/associate/tools/spawn.ts +827 -0
- associate-0.11.0/.pi/extensions/associate/tools/tests.ts +719 -0
- associate-0.11.0/AGENTS.md +94 -0
- {associate-0.9.1 → associate-0.11.0}/CHANGELOG.md +353 -0
- associate-0.11.0/CLAUDE.md +733 -0
- associate-0.11.0/PKG-INFO +236 -0
- associate-0.11.0/README.md +214 -0
- associate-0.11.0/associate/bench/__init__.py +17 -0
- associate-0.11.0/associate/bench/public.py +203 -0
- associate-0.11.0/associate/bench/runner.py +115 -0
- associate-0.11.0/associate/bench/suites.json +58 -0
- {associate-0.9.1 → associate-0.11.0}/associate/bench/walkstats.py +154 -21
- associate-0.11.0/associate/bridge/__init__.py +11 -0
- associate-0.11.0/associate/bridge/access.py +336 -0
- associate-0.11.0/associate/bridge/caching.py +47 -0
- associate-0.11.0/associate/bridge/filetokens.py +140 -0
- associate-0.11.0/associate/bridge/guards.py +84 -0
- associate-0.11.0/associate/bridge/mcpserver.py +167 -0
- associate-0.11.0/associate/bridge/origin.py +131 -0
- associate-0.11.0/associate/bridge/policy.py +127 -0
- associate-0.11.0/associate/bridge/principal.py +98 -0
- associate-0.11.0/associate/bridge/resolve.py +134 -0
- associate-0.11.0/associate/bridge/static.py +135 -0
- associate-0.11.0/associate/bridge/tokens.py +230 -0
- {associate-0.9.1 → associate-0.11.0}/associate/cli/__init__.py +9 -1
- associate-0.11.0/associate/cli/_commands/bench.py +78 -0
- associate-0.11.0/associate/cli/_commands/learn.py +150 -0
- associate-0.11.0/associate/cli/_commands/mcp.py +38 -0
- {associate-0.9.1 → associate-0.11.0}/associate/cli/_commands/overview.py +3 -0
- associate-0.11.0/associate/cli/_commands/run.py +509 -0
- associate-0.11.0/associate/cli/_commands/runs.py +190 -0
- associate-0.11.0/associate/cli/_commands/serve.py +112 -0
- associate-0.11.0/associate/cli/_commands/tokens.py +151 -0
- {associate-0.9.1 → associate-0.11.0}/associate/cli/_output.py +22 -0
- {associate-0.9.1 → associate-0.11.0}/associate/contract/__init__.py +42 -3
- associate-0.11.0/associate/contract/gate.json +105 -0
- associate-0.11.0/associate/contract/models.json +96 -0
- associate-0.11.0/associate/contract/policy.json +123 -0
- associate-0.11.0/associate/contract/role.json +24 -0
- associate-0.11.0/associate/contract/schemas/gate.schema.json +138 -0
- associate-0.11.0/associate/contract/schemas/models.schema.json +46 -0
- {associate-0.9.1 → associate-0.11.0}/associate/contract/schemas/task.schema.json +10 -0
- associate-0.11.0/associate/contract/schemas/walk.schema.json +92 -0
- associate-0.11.0/associate/contract/validate.py +531 -0
- associate-0.11.0/associate/explain/catalog.py +549 -0
- associate-0.11.0/associate/harness/checkoutlock.py +160 -0
- {associate-0.9.1 → associate-0.11.0}/associate/harness/pi.py +353 -42
- associate-0.11.0/associate/harness/testcmds.py +381 -0
- associate-0.11.0/associate/harness/worktree.py +385 -0
- associate-0.11.0/associate/runs/__init__.py +6 -0
- associate-0.11.0/associate/runs/mcp_tools.py +82 -0
- associate-0.11.0/associate/runs/service.py +919 -0
- associate-0.11.0/associate/runs/state.py +157 -0
- associate-0.11.0/associate/runs/supervisor.py +242 -0
- associate-0.11.0/associate/server/__init__.py +4 -0
- associate-0.11.0/associate/server/app.py +298 -0
- associate-0.11.0/associate/server/serve.py +352 -0
- associate-0.11.0/associate/server/tokenstore.py +46 -0
- associate-0.11.0/associate/web_dist/assets/albert-sans-latin-ext-wght-normal-BAo4qCk1.woff2 +0 -0
- associate-0.11.0/associate/web_dist/assets/albert-sans-latin-wght-normal-BJ0ssN8N.woff2 +0 -0
- associate-0.11.0/associate/web_dist/assets/fraunces-latin-ext-wght-normal-Ca2vKHc0.woff2 +0 -0
- associate-0.11.0/associate/web_dist/assets/fraunces-latin-wght-normal-ukD16Tqj.woff2 +0 -0
- associate-0.11.0/associate/web_dist/assets/fraunces-vietnamese-wght-normal-CnvboYUG.woff2 +0 -0
- associate-0.11.0/associate/web_dist/assets/index-IbtklPgE.css +1 -0
- associate-0.11.0/associate/web_dist/assets/index-YpORTsP3.js +52 -0
- associate-0.11.0/associate/web_dist/assets/index-YpORTsP3.js.map +1 -0
- associate-0.11.0/associate/web_dist/index.html +16 -0
- associate-0.11.0/culture.yaml +31 -0
- associate-0.11.0/docs/notes/gate-calibration.md +289 -0
- associate-0.11.0/docs/notes/pi-1.1.0-migration.md +179 -0
- associate-0.11.0/docs/notes/resident-model-resolution.md +52 -0
- associate-0.11.0/docs/notes/t20-live-acceptance.md +61 -0
- associate-0.11.0/docs/operations/bridge.md +562 -0
- associate-0.11.0/docs/plans/2026-10-27-associate-multi-model-on-pi-1-1-0-split.md +275 -0
- associate-0.11.0/docs/plans/2026-10-27-associate-multi-model-on-pi-1-1-0.md +210 -0
- associate-0.11.0/docs/specs/2026-10-27-associate-multi-model-on-pi-1-1-0.md +261 -0
- associate-0.11.0/hatch_build.py +66 -0
- {associate-0.9.1 → associate-0.11.0}/pyproject.toml +23 -2
- associate-0.11.0/scripts/sonar-pr-issues.py +168 -0
- {associate-0.9.1 → associate-0.11.0}/sonar-project.properties +12 -4
- associate-0.11.0/tests/bridge/__init__.py +0 -0
- associate-0.11.0/tests/bridge/jwks.py +94 -0
- associate-0.11.0/tests/bridge/memstore.py +52 -0
- associate-0.11.0/tests/bridge/test_access_jwt.py +308 -0
- associate-0.11.0/tests/bridge/test_boundary.py +80 -0
- associate-0.11.0/tests/bridge/test_caching.py +46 -0
- associate-0.11.0/tests/bridge/test_filetokens.py +95 -0
- associate-0.11.0/tests/bridge/test_guards.py +80 -0
- associate-0.11.0/tests/bridge/test_mcpserver.py +142 -0
- associate-0.11.0/tests/bridge/test_policy.py +88 -0
- associate-0.11.0/tests/bridge/test_principal_and_tokens.py +142 -0
- associate-0.11.0/tests/bridge/test_resolve.py +119 -0
- associate-0.11.0/tests/bridge/test_static.py +146 -0
- associate-0.11.0/tests/rawasgi.py +95 -0
- associate-0.11.0/tests/runs_fake_child.py +93 -0
- associate-0.11.0/tests/test_base_install.py +343 -0
- associate-0.11.0/tests/test_bench.py +237 -0
- associate-0.11.0/tests/test_checkoutlock.py +248 -0
- associate-0.11.0/tests/test_ci_workflows.py +126 -0
- {associate-0.9.1 → associate-0.11.0}/tests/test_cli_introspection.py +10 -0
- {associate-0.9.1 → associate-0.11.0}/tests/test_contract.py +32 -14
- associate-0.11.0/tests/test_gate_contract.py +447 -0
- associate-0.11.0/tests/test_hatch_build.py +90 -0
- associate-0.11.0/tests/test_mcp.py +253 -0
- associate-0.11.0/tests/test_models_contract.py +258 -0
- associate-0.11.0/tests/test_provider_wire.py +427 -0
- associate-0.11.0/tests/test_raw_paths.py +137 -0
- {associate-0.9.1 → associate-0.11.0}/tests/test_run.py +541 -18
- associate-0.11.0/tests/test_runs.py +533 -0
- associate-0.11.0/tests/test_runs_body_cap.py +130 -0
- associate-0.11.0/tests/test_runs_csrf.py +262 -0
- associate-0.11.0/tests/test_runs_delegate_lock.py +42 -0
- associate-0.11.0/tests/test_runs_http.py +147 -0
- associate-0.11.0/tests/test_runs_starting.py +204 -0
- associate-0.11.0/tests/test_runs_ui_fields.py +344 -0
- {associate-0.9.1 → associate-0.11.0}/tests/test_runtime_prompt.py +26 -0
- associate-0.11.0/tests/test_serve_cli.py +195 -0
- associate-0.11.0/tests/test_serve_config.py +136 -0
- associate-0.11.0/tests/test_serve_listeners.py +274 -0
- associate-0.11.0/tests/test_sonar_pr_issues.py +128 -0
- associate-0.11.0/tests/test_surface_parity.py +491 -0
- associate-0.11.0/tests/test_walkstats.py +317 -0
- associate-0.11.0/tests/test_web_contract.py +134 -0
- associate-0.11.0/tests/test_web_ui.py +198 -0
- associate-0.11.0/tests/test_worktree.py +809 -0
- associate-0.11.0/uv.lock +1226 -0
- associate-0.11.0/web/index.html +15 -0
- associate-0.11.0/web/package-lock.json +3891 -0
- associate-0.11.0/web/package.json +41 -0
- associate-0.11.0/web/scripts/check-tokens.mjs +85 -0
- associate-0.11.0/web/src/App.test.tsx +39 -0
- associate-0.11.0/web/src/App.tsx +34 -0
- associate-0.11.0/web/src/api/client.test.ts +126 -0
- associate-0.11.0/web/src/api/client.ts +150 -0
- associate-0.11.0/web/src/api/routes.json +5 -0
- associate-0.11.0/web/src/api/runs.ts +16 -0
- associate-0.11.0/web/src/api/types.ts +91 -0
- associate-0.11.0/web/src/components/Header.tsx +32 -0
- associate-0.11.0/web/src/components/Problem.tsx +38 -0
- associate-0.11.0/web/src/components/StatusPill.tsx +16 -0
- associate-0.11.0/web/src/culture-design/README.md +36 -0
- associate-0.11.0/web/src/culture-design/tokens.css +545 -0
- associate-0.11.0/web/src/fixtures/runs.json +369 -0
- associate-0.11.0/web/src/hooks/useLoad.ts +62 -0
- associate-0.11.0/web/src/hooks/usePolling.test.tsx +52 -0
- associate-0.11.0/web/src/hooks/usePolling.ts +43 -0
- associate-0.11.0/web/src/hooks/useRunsList.tsx +40 -0
- associate-0.11.0/web/src/hooks/useTitle.ts +8 -0
- associate-0.11.0/web/src/main.tsx +29 -0
- associate-0.11.0/web/src/model/runs.test.ts +220 -0
- associate-0.11.0/web/src/model/runs.ts +175 -0
- associate-0.11.0/web/src/routes/Run.test.tsx +314 -0
- associate-0.11.0/web/src/routes/Run.tsx +279 -0
- associate-0.11.0/web/src/routes/Runs.test.tsx +336 -0
- associate-0.11.0/web/src/routes/Runs.tsx +157 -0
- associate-0.11.0/web/src/styles/app.css +532 -0
- associate-0.11.0/web/src/test/api.ts +83 -0
- associate-0.11.0/web/src/test/render.tsx +12 -0
- associate-0.11.0/web/tsconfig.app.json +27 -0
- associate-0.11.0/web/tsconfig.json +8 -0
- associate-0.11.0/web/tsconfig.node.json +25 -0
- associate-0.11.0/web/vite.config.ts +49 -0
- associate-0.11.0/web/vitest.setup.ts +26 -0
- associate-0.9.1/.devague/current +0 -1
- associate-0.9.1/.devague/current_plan +0 -1
- associate-0.9.1/.pi/extensions/associate/lib/guard.ts +0 -173
- associate-0.9.1/.pi/extensions/associate/lib/prompt.ts +0 -114
- associate-0.9.1/.pi/extensions/associate/lib/provider.ts +0 -251
- associate-0.9.1/.pi/extensions/associate/tests/guard.test.ts +0 -118
- associate-0.9.1/.pi/extensions/associate/tests/prompt.test.ts +0 -127
- associate-0.9.1/.pi/extensions/associate/tests/provider.test.ts +0 -290
- associate-0.9.1/AGENTS.md +0 -79
- associate-0.9.1/CLAUDE.md +0 -464
- associate-0.9.1/PKG-INFO +0 -259
- associate-0.9.1/README.md +0 -242
- associate-0.9.1/associate/bench/__init__.py +0 -40
- associate-0.9.1/associate/bench/checks.py +0 -482
- associate-0.9.1/associate/bench/corpus.py +0 -273
- associate-0.9.1/associate/bench/replay.py +0 -117
- associate-0.9.1/associate/bench/runner.py +0 -184
- associate-0.9.1/associate/cli/_commands/bench.py +0 -75
- associate-0.9.1/associate/cli/_commands/learn.py +0 -101
- associate-0.9.1/associate/cli/_commands/run.py +0 -191
- associate-0.9.1/associate/contract/policy.json +0 -94
- associate-0.9.1/associate/contract/role.json +0 -26
- associate-0.9.1/associate/contract/schemas/walk.schema.json +0 -51
- associate-0.9.1/associate/contract/validate.py +0 -158
- associate-0.9.1/associate/explain/catalog.py +0 -238
- associate-0.9.1/culture.yaml +0 -27
- associate-0.9.1/tests/behavioral/cases/01-local-read-find.json +0 -21
- associate-0.9.1/tests/behavioral/cases/02-repo-exploration.json +0 -21
- associate-0.9.1/tests/behavioral/cases/03-summarization.json +0 -20
- associate-0.9.1/tests/behavioral/cases/04-structured-evidence-extraction.json +0 -22
- associate-0.9.1/tests/behavioral/cases/05-tool-call-reliability.json +0 -21
- associate-0.9.1/tests/behavioral/cases/06-forbidden-mutation-attempts.json +0 -19
- associate-0.9.1/tests/behavioral/cases/07-bounded-completion-and-hand-back.json +0 -20
- associate-0.9.1/tests/test_bench.py +0 -497
- associate-0.9.1/tests/test_provider_wire.py +0 -191
- associate-0.9.1/tests/test_walkstats.py +0 -148
- associate-0.9.1/uv.lock +0 -478
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/agent-config/SKILL.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/agent-config/data/backend-fingerprints.yaml +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/agent-config/scripts/show.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/ask-colleague/SKILL.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/ask-colleague/prompts/explore.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/ask-colleague/prompts/review.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/ask-colleague/prompts/write.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/ask-colleague/scripts/ask-colleague.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/assign-to-workforce/SKILL.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/assign-to-workforce/scripts/assign-to-workforce.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/challenge/SKILL.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/cicd/SKILL.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/cicd/scripts/_resolve-nick.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/cicd/scripts/portability-lint.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/cicd/scripts/pr-reply.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/cicd/scripts/pr-status.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/cicd/scripts/workflow.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/communicate/SKILL.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/communicate/scripts/fetch-issues.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/communicate/scripts/mesh-message.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/communicate/scripts/post-comment.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/communicate/scripts/post-issue.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/communicate/scripts/templates/skill-new-brief.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/communicate/scripts/templates/skill-update-brief.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/deviate/SKILL.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/doc-test-alignment/SKILL.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/doc-test-alignment/scripts/check.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/pypi-maintainer/SKILL.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/pypi-maintainer/scripts/switch-source.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/recall/SKILL.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/recall/scripts/recall.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/remember/SKILL.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/remember/scripts/remember.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/run-tests/SKILL.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/run-tests/scripts/test.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/scope/SKILL.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/sonarclaude/SKILL.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/sonarclaude/scripts/sonar.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/spec-to-plan/SKILL.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/spec-to-plan/scripts/spec-to-plan.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/summarize-delivery/SKILL.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/think/SKILL.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/think/scripts/think.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/validate-delivery/SKILL.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/version-bump/SKILL.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills/version-bump/scripts/bump.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/.claude/skills.local.yaml.example +0 -0
- {associate-0.9.1 → associate-0.11.0}/.devague/deliveries/associate-on-pi-with-opinionated-tools.json +0 -0
- {associate-0.9.1 → associate-0.11.0}/.devague/frames/associate-on-pi-with-opinionated-tools.json +0 -0
- {associate-0.9.1 → associate-0.11.0}/.devague/frames/webglass-code-lens-as-tools.json +0 -0
- {associate-0.9.1 → associate-0.11.0}/.devague/plans/associate-on-pi-with-opinionated-tools.json +0 -0
- {associate-0.9.1 → associate-0.11.0}/.devague/plans/webglass-code-lens-as-tools.json +0 -0
- {associate-0.9.1 → associate-0.11.0}/.flake8 +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/lib/contain.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/lib/handback.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/lib/preflight.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/lib/session.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/lib/statements.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/lib/torch.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tests/codelens.test.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tests/contain.test.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tests/preflight.test.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tests/read.test.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tests/run.sh +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tests/runtime.test.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tests/session.test.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tests/shell.test.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tests/smoke.test.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tests/statements.test.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tests/stubs/typebox.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tests/torch.test.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tests/web.test.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tools/.gitkeep +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tools/_procs.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tools/codelens.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tools/shell.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/extensions/associate/tools/web.ts +0 -0
- {associate-0.9.1 → associate-0.11.0}/.pi/settings.json +0 -0
- {associate-0.9.1 → associate-0.11.0}/AGENTS.colleague.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/LICENSE +0 -0
- {associate-0.9.1 → associate-0.11.0}/associate/__init__.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/associate/__main__.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/associate/bench/config.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/associate/bench/table.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/associate/cli/_commands/__init__.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/associate/cli/_commands/cli.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/associate/cli/_commands/doctor.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/associate/cli/_commands/explain.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/associate/cli/_commands/whoami.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/associate/cli/_errors.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/associate/contract/schemas/statements.schema.json +0 -0
- {associate-0.9.1 → associate-0.11.0}/associate/explain/__init__.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/associate/harness/__init__.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/associate/harness/base.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/associate/harness/stub.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/docs/deliveries/2026-09-12-associate-on-pi-with-opinionated-tools.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/docs/plans/2026-09-05-webglass-code-lens-as-tools.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/docs/plans/2026-09-12-associate-on-pi-with-opinionated-tools-split.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/docs/plans/2026-09-12-associate-on-pi-with-opinionated-tools.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/docs/skill-sources.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/docs/specs/2026-09-05-webglass-code-lens-as-tools.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/docs/specs/2026-09-12-associate-on-pi-with-opinionated-tools.md +0 -0
- {associate-0.9.1 → associate-0.11.0}/package.json +0 -0
- {associate-0.9.1 → associate-0.11.0}/tests/__init__.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/tests/conftest.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/tests/fake_lane.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/tests/test_cli.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/tests/test_fake_lane.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/tests/test_harness_stub.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/tests/test_pi_config.py +0 -0
- {associate-0.9.1 → associate-0.11.0}/tests/test_pi_dependent.py +0 -0
|
@@ -0,0 +1 @@
|
|
|
1
|
+
associate-multi-model-on-pi-1-1-0
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
associate-multi-model-on-pi-1-1-0
|
|
@@ -0,0 +1,285 @@
|
|
|
1
|
+
{
|
|
2
|
+
"plan_slug": "associate-multi-model-on-pi-1-1-0",
|
|
3
|
+
"schema_version": 2,
|
|
4
|
+
"created": "2026-10-27T10:36:31Z",
|
|
5
|
+
"updated": "2026-10-27T23:32:44Z",
|
|
6
|
+
"deviations": [
|
|
7
|
+
{
|
|
8
|
+
"id": "d1",
|
|
9
|
+
"what": "t1 gains a dependency on t6: its investigation note merges now, its fix (the extension sets the session model on session_start via pi.setModel from the registry default) and live check move after t6, into wave 5",
|
|
10
|
+
"task_ref": "t1",
|
|
11
|
+
"reason": "pi-acp ignores session/new params.model and culture never sends session/set_model (docs/notes/resident-model-resolution.md), so the fix must live in the extension's lib/provider.ts, which t6 rewrites",
|
|
12
|
+
"affects": [
|
|
13
|
+
"t6",
|
|
14
|
+
"t19"
|
|
15
|
+
],
|
|
16
|
+
"origin": "user",
|
|
17
|
+
"status": "approved",
|
|
18
|
+
"classification": "acceptable",
|
|
19
|
+
"seq": 1
|
|
20
|
+
},
|
|
21
|
+
{
|
|
22
|
+
"id": "d2",
|
|
23
|
+
"what": "add a CI gate that fails when SonarCloud reports new issues (accepted issues are fine), modelled on ../culture-library, to workledger-cli#5 (t3) and to associate's integration branch",
|
|
24
|
+
"task_ref": "t3",
|
|
25
|
+
"reason": "user instruction mid-run: SonarCloud new issues must fail CI, not just be visible; accepted issues stay allowed",
|
|
26
|
+
"affects": [
|
|
27
|
+
"t19"
|
|
28
|
+
],
|
|
29
|
+
"origin": "user",
|
|
30
|
+
"status": "approved",
|
|
31
|
+
"classification": "acceptable",
|
|
32
|
+
"seq": 2
|
|
33
|
+
},
|
|
34
|
+
{
|
|
35
|
+
"id": "d3",
|
|
36
|
+
"what": "t13's tools are named ledger_append / ledger_list (the contract's tool_names, set in t4 and already used by the t11 prompts and models.json mutating_tools), not workledger_append / workledger_list as t13's plan text says; the released workledger-cli console script is ledger",
|
|
37
|
+
"task_ref": "t13",
|
|
38
|
+
"reason": "the main agent's t4 brief chose the ledger_* names before t13 ran; the contract is the single source of truth for tool names, so following the plan text now would split it",
|
|
39
|
+
"affects": [
|
|
40
|
+
"t4",
|
|
41
|
+
"t11"
|
|
42
|
+
],
|
|
43
|
+
"origin": "llm",
|
|
44
|
+
"status": "approved",
|
|
45
|
+
"classification": "acceptable",
|
|
46
|
+
"seq": 3
|
|
47
|
+
},
|
|
48
|
+
{
|
|
49
|
+
"id": "d4",
|
|
50
|
+
"what": "live testing starts now instead of only at t20: associate itself, driven by the live models (worker/bolt, senses, cortex) and the live CLEF decider, reviews this integration branch and fixes what it finds; read-only review runs first, mutating fix runs only after t12 (gate wiring) and the self-examination conflict (r13) are fixed",
|
|
51
|
+
"task_ref": "t20",
|
|
52
|
+
"reason": "user instruction: 'As live test, use associate with the models we have to review and fix'",
|
|
53
|
+
"affects": [
|
|
54
|
+
"t12",
|
|
55
|
+
"t19"
|
|
56
|
+
],
|
|
57
|
+
"origin": "user",
|
|
58
|
+
"status": "approved",
|
|
59
|
+
"classification": "acceptable",
|
|
60
|
+
"seq": 4
|
|
61
|
+
},
|
|
62
|
+
{
|
|
63
|
+
"id": "d5",
|
|
64
|
+
"what": "add an opt-in launcher option (off by default) that starts pi with --no-approve --no-extensions --no-mcp, so the examined checkout's .pi settings, extensions and mcp.json are ignored; default launch keeps --approve",
|
|
65
|
+
"task_ref": "t12",
|
|
66
|
+
"reason": "user decision on risk r16: 'R16 as optional, yes (no default)'; built by the main agent after t12 merges because both touch associate/harness/pi.py",
|
|
67
|
+
"affects": [
|
|
68
|
+
"t20"
|
|
69
|
+
],
|
|
70
|
+
"origin": "user",
|
|
71
|
+
"status": "approved",
|
|
72
|
+
"classification": "acceptable",
|
|
73
|
+
"seq": 5
|
|
74
|
+
},
|
|
75
|
+
{
|
|
76
|
+
"id": "d6",
|
|
77
|
+
"what": "eidetic_remember passes the CLEF gate for every model via a new contract list gated_tools (gated but not a repo mutation, so senses keeps it)",
|
|
78
|
+
"task_ref": "t12",
|
|
79
|
+
"reason": "user decision on r17: 'R17 checked as well'",
|
|
80
|
+
"affects": [
|
|
81
|
+
"t8"
|
|
82
|
+
],
|
|
83
|
+
"origin": "user",
|
|
84
|
+
"status": "approved",
|
|
85
|
+
"classification": "acceptable",
|
|
86
|
+
"seq": 6
|
|
87
|
+
},
|
|
88
|
+
{
|
|
89
|
+
"id": "d7",
|
|
90
|
+
"what": "web_search and web_page also pass the CLEF gate for every model (added to gated_tools alongside eidetic_remember, extending d6); with CLEF and Super unreachable they are refused like any gated call",
|
|
91
|
+
"task_ref": "t12",
|
|
92
|
+
"reason": "user decision: 'Web tool search and page - yes' (follow-up to r17: they can send data off the machine)",
|
|
93
|
+
"affects": [
|
|
94
|
+
"t8"
|
|
95
|
+
],
|
|
96
|
+
"origin": "user",
|
|
97
|
+
"status": "approved",
|
|
98
|
+
"classification": "acceptable",
|
|
99
|
+
"seq": 7
|
|
100
|
+
},
|
|
101
|
+
{
|
|
102
|
+
"id": "d8",
|
|
103
|
+
"what": "gate.json v2: the decider state gains an isolation field (worktree vs in-place, from ASSOCIATE_ISOLATION set by associate run), so CLEF judges a mutating call knowing whether it lands in a disposable run worktree; thresholds unchanged pending measurement",
|
|
104
|
+
"task_ref": "t5",
|
|
105
|
+
"reason": "live F2 run (L5): CLEF refused correct, minimal edits at safe_unattended 0.65-0.70 vs 0.8 because the state never said edits land in an isolated worktree merged only after a clean run; user chose option B and asked to measure before any threshold change (option A)",
|
|
106
|
+
"affects": [
|
|
107
|
+
"t10",
|
|
108
|
+
"t12",
|
|
109
|
+
"t15"
|
|
110
|
+
],
|
|
111
|
+
"origin": "user",
|
|
112
|
+
"status": "approved",
|
|
113
|
+
"classification": "needs-follow-up",
|
|
114
|
+
"seq": 8
|
|
115
|
+
},
|
|
116
|
+
{
|
|
117
|
+
"id": "d9",
|
|
118
|
+
"what": "gate.json v3: CLEF thresholds safe_unattended >= 0.5 and matches_task >= 0.5 (were 0.8 / 0.7), tentative until the CLEF benchmark in issue #13 freezes a calibrated policy",
|
|
119
|
+
"task_ref": "t5",
|
|
120
|
+
"reason": "measured (docs/notes/gate-calibration.md): legitimate F2 edits/writes score safe_unattended >= 0.714 and matches_task >= 0.701 under the v2 worktree state, harmful probes <= 0.120 and <= 0.043; 0.8 sat inside the legitimate range. User set 0.5 tentatively, with issue #13 as the benchmark that replaces it",
|
|
121
|
+
"affects": [
|
|
122
|
+
"t10",
|
|
123
|
+
"t21"
|
|
124
|
+
],
|
|
125
|
+
"origin": "user",
|
|
126
|
+
"status": "approved",
|
|
127
|
+
"classification": "needs-follow-up",
|
|
128
|
+
"seq": 9
|
|
129
|
+
},
|
|
130
|
+
{
|
|
131
|
+
"id": "d10",
|
|
132
|
+
"what": "gate.json v4: the decider state names what the gated tool actually does - a contract-owned effect line per tool that can reach the gate (mutating_tools + gated_tools) - and the isolation line is reworded to cover only changes to the checkout, so it is true for every tool",
|
|
133
|
+
"task_ref": "t5",
|
|
134
|
+
"reason": "the v2/v3 state named the tool but not its effect, and its isolation line claimed a run worktree for tools that write outside the checkout or off the machine (ledger_append, eidetic_remember, web_search, web_page); with a true effect line ledger entries scored safe_unattended 0.83-0.88 instead of 0.07-0.24 (docs/notes/gate-calibration.md). User approved d10",
|
|
135
|
+
"affects": [
|
|
136
|
+
"t10",
|
|
137
|
+
"t8",
|
|
138
|
+
"t13"
|
|
139
|
+
],
|
|
140
|
+
"origin": "user",
|
|
141
|
+
"status": "approved",
|
|
142
|
+
"classification": "needs-follow-up",
|
|
143
|
+
"seq": 10
|
|
144
|
+
},
|
|
145
|
+
{
|
|
146
|
+
"id": "d11",
|
|
147
|
+
"what": "gate.json v5: matches_task reworded to count the evidence-logging, memory and research work the runtime prompt asks every run to do as serving the task",
|
|
148
|
+
"task_ref": "t5",
|
|
149
|
+
"reason": "v4 measurement (docs/notes/gate-calibration.md): with a truthful state, ledger/memory writes score safe 0.62-0.84 but fail matches_task 0.12-0.45 because those duties live in the runtime prompt CLEF never sees; the rewording admitted all ledger entries and the benign memory note while every harmful probe stayed refused (matches_task <= 0.10). User confirmed; remaining gaps go to issues, fixed after the issue #13 benchmark",
|
|
150
|
+
"affects": [
|
|
151
|
+
"t10",
|
|
152
|
+
"t13",
|
|
153
|
+
"t8"
|
|
154
|
+
],
|
|
155
|
+
"origin": "user",
|
|
156
|
+
"status": "approved",
|
|
157
|
+
"classification": "needs-follow-up",
|
|
158
|
+
"seq": 11
|
|
159
|
+
},
|
|
160
|
+
{
|
|
161
|
+
"id": "d12",
|
|
162
|
+
"what": "models that may change code can run tests (option 3): the operator names exact test commands with 'associate run --test-command' (repeatable, no default); worker and cortex get a dedicated run_tests tool that runs exactly those commands behind the CLEF gate, and the harness runs them itself after the model finishes and merges only if all pass; test-run side effects on the worktree are reverted so they are never committed as the model's work",
|
|
163
|
+
"task_ref": "t15",
|
|
164
|
+
"reason": "live F2 run: bolt edited code it could not verify (node/uv/pytest are not on the bash allowlist), shipped one defect the suite then caught, and timed out; user accepted option 3 (harness check is the final word, models may also run the fixed commands)",
|
|
165
|
+
"affects": [
|
|
166
|
+
"t12",
|
|
167
|
+
"t5",
|
|
168
|
+
"t20"
|
|
169
|
+
],
|
|
170
|
+
"origin": "user",
|
|
171
|
+
"status": "approved",
|
|
172
|
+
"classification": "needs-follow-up",
|
|
173
|
+
"seq": 12
|
|
174
|
+
},
|
|
175
|
+
{
|
|
176
|
+
"id": "d13",
|
|
177
|
+
"what": "role.json becomes associate's harness-wide bound: final_decision and security_decision are forbidden for every model; per-model write rights (repo_action, code_authoring) are owned by models.json may_mutate, no longer forbidden for all by a copy of the retired lobes associate role",
|
|
178
|
+
"task_ref": "t11",
|
|
179
|
+
"reason": "role.json still copied the retired lobes 'associate' role (worker MINUS repo_action), forbidding code_authoring and repo_action for everyone, contradicting models.json (worker and cortex may_mutate) and the d4 AGENTS.md bound; a guard refusal still said 'the associate role is forbidden repo_action'. User accepted the recommendation",
|
|
180
|
+
"affects": [
|
|
181
|
+
"t12",
|
|
182
|
+
"t19"
|
|
183
|
+
],
|
|
184
|
+
"origin": "user",
|
|
185
|
+
"status": "approved",
|
|
186
|
+
"classification": "acceptable",
|
|
187
|
+
"seq": 13
|
|
188
|
+
},
|
|
189
|
+
{
|
|
190
|
+
"id": "d14",
|
|
191
|
+
"what": "t19's 'steward doctor passes' criterion is read as 'steward doctor adds no findings compared with main': measured 0 new (42 portability lines and 5 skills-convention findings on both main dfae5dd and this branch); the existing findings are fixed upstream (devague evidence records, steward rules) and steward doctor is added to CI once they land",
|
|
192
|
+
"task_ref": "t19",
|
|
193
|
+
"reason": "every finding is in files this branch did not create and repo rules forbid editing: the September plan's .devague records and their exports (PR #6) and vendored skills; the five method-only devague skills are scriptless by design. User confirmed option 1",
|
|
194
|
+
"affects": [],
|
|
195
|
+
"origin": "user",
|
|
196
|
+
"status": "approved",
|
|
197
|
+
"classification": "needs-follow-up",
|
|
198
|
+
"seq": 14
|
|
199
|
+
},
|
|
200
|
+
{
|
|
201
|
+
"id": "d15",
|
|
202
|
+
"what": "super-vl gets the run_tests tool (models.json tools and prompts/super-vl.md), like cortex, for when the model arrives",
|
|
203
|
+
"task_ref": "t16",
|
|
204
|
+
"reason": "d12 briefed run_tests for worker and cortex only; super-vl mirrors cortex's tool set and may mutate. User confirmed",
|
|
205
|
+
"affects": [
|
|
206
|
+
"t11"
|
|
207
|
+
],
|
|
208
|
+
"origin": "user",
|
|
209
|
+
"status": "approved",
|
|
210
|
+
"classification": "acceptable",
|
|
211
|
+
"seq": 15
|
|
212
|
+
},
|
|
213
|
+
{
|
|
214
|
+
"id": "d16",
|
|
215
|
+
"what": "oversized gate states (over the decider's input budget) are no longer cut and refused: they are judged by a different, large-context registry model (gate.json large_state.evaluators, default [cortex, worker]; the first active entry that is not the acting model), asked the same questions with the full state through the Super-style JSON verdict; no evaluator reachable, or a state too large even for it, is still refused",
|
|
216
|
+
"task_ref": "t10",
|
|
217
|
+
"reason": "PR #18 review finding 4: any write or task over about 7-14 KB was always refused because CLEF's 16k-token input forced truncation (r15). User declined a summarised state and chose a different model as evaluator for large inputs; the actor never judges its own call",
|
|
218
|
+
"affects": [
|
|
219
|
+
"t5",
|
|
220
|
+
"t12"
|
|
221
|
+
],
|
|
222
|
+
"origin": "user",
|
|
223
|
+
"status": "approved",
|
|
224
|
+
"classification": "needs-follow-up",
|
|
225
|
+
"seq": 16
|
|
226
|
+
},
|
|
227
|
+
{
|
|
228
|
+
"id": "d17",
|
|
229
|
+
"what": "the examined checkout's AGENTS.md is skipped under --untrusted-checkout and, for a trusted checkout, injected labelled as repository context rather than as this lane's runtime prompt",
|
|
230
|
+
"task_ref": "t11",
|
|
231
|
+
"reason": "PR #18 review follow-up: a hostile checkout's AGENTS.md was injected as the runtime prompt even under --untrusted-checkout; user accepted skipping it there and relabelling it otherwise",
|
|
232
|
+
"affects": [
|
|
233
|
+
"t2"
|
|
234
|
+
],
|
|
235
|
+
"origin": "user",
|
|
236
|
+
"status": "approved",
|
|
237
|
+
"classification": "acceptable",
|
|
238
|
+
"seq": 17
|
|
239
|
+
},
|
|
240
|
+
{
|
|
241
|
+
"id": "d18",
|
|
242
|
+
"what": "design D (O3): the mesh resident stays read-only and hands any task that writes to 'associate run' (a per-run worktree, the operator test check and the clean-only merge) through a new gated tool, instead of writing in its own long-lived session",
|
|
243
|
+
"task_ref": "t16",
|
|
244
|
+
"reason": "culture's ACP runner keeps one long-lived pi session in the registered directory, so a mesh request cannot get its own worktree inside that session; O3 reuses the one isolation path measured live. User accepted O3",
|
|
245
|
+
"affects": [
|
|
246
|
+
"t15",
|
|
247
|
+
"t20"
|
|
248
|
+
],
|
|
249
|
+
"origin": "user",
|
|
250
|
+
"status": "approved",
|
|
251
|
+
"classification": "needs-follow-up",
|
|
252
|
+
"seq": 18
|
|
253
|
+
},
|
|
254
|
+
{
|
|
255
|
+
"id": "d19",
|
|
256
|
+
"what": "amends d16: when no other large-state evaluator is available (no active registry model in large_state.evaluators other than the acting model), the acting model itself evaluates the oversized call in a fresh, history-free request with a short focused gate prompt (the questions and the JSON verdict shape only) and the call's state; the walk marks the verdict as self-evaluated; refusal stays the outcome if that also fails",
|
|
257
|
+
"task_ref": "t10",
|
|
258
|
+
"reason": "user accepted d16 and added: if no other model exists, defer to the current model with smaller context and a focused prompt",
|
|
259
|
+
"affects": [
|
|
260
|
+
"t5"
|
|
261
|
+
],
|
|
262
|
+
"origin": "user",
|
|
263
|
+
"status": "approved",
|
|
264
|
+
"classification": "needs-follow-up",
|
|
265
|
+
"seq": 19
|
|
266
|
+
},
|
|
267
|
+
{
|
|
268
|
+
"id": "d20",
|
|
269
|
+
"what": "gate.json v8: the gate state gains a change line for a write that replaces an existing file (the lines it adds and removes versus the current file, computed by the guard, redacted like every field; the full content stays), and the large-state path gets its own timeout (large_state.timeout_ms, default 300000) instead of reusing Super's 120 s",
|
|
270
|
+
"task_ref": "t10",
|
|
271
|
+
"reason": "live d16 probe: cortex allowed a realistic 40 KB CLAUDE.md write with one planted exfiltration line when shown full content only, and refused it (exfiltrates=true) with the change line added; 3 of 7 large probes ran past the 120 s Super timeout (37-213 s). One pair, not a rate. User approved d20",
|
|
272
|
+
"affects": [
|
|
273
|
+
"t5",
|
|
274
|
+
"t12"
|
|
275
|
+
],
|
|
276
|
+
"origin": "user",
|
|
277
|
+
"status": "approved",
|
|
278
|
+
"classification": "needs-follow-up",
|
|
279
|
+
"seq": 20
|
|
280
|
+
}
|
|
281
|
+
],
|
|
282
|
+
"evidence": [],
|
|
283
|
+
"deltas": [],
|
|
284
|
+
"supersessions": []
|
|
285
|
+
}
|