Commit Graph
6670 Commits
Author SHA1 Message Date
Dotta 1ad2507b85 fix(evals): distinguish timeline actors from deadline adjuncts 2026-11-04 14:13:29 -06:00
Dotta 9f8394c532 fix(evals): bound timeline object and result assertions 2026-11-04 14:13:29 -06:00
Dotta 58ff18c731 fix(evals): validate timeline actor clauses and exit outcomes 2026-11-04 13:46:54 -06:00
Dotta 1b5ebfaf7b fix(evals): share timeline takeover provenance guards 2026-11-04 13:40:12 -06:00
Dotta 3a9c3fa8cc fix(evals): recognize successor takeover in timeline context 2026-11-04 13:40:12 -06:00
Dotta 2569e44944 fix(evals): accept recorded passive timeline handoffs 2026-11-04 13:40:12 -06:00
Dotta b20730d647 fix(server): clarify failed-action chat completion feedback 2026-11-04 13:36:09 -06:00
DottaandPaperclip 5c378d5837 test(evals): version all credential destination consumers
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 13:27:42 -06:00
DottaandPaperclip 8da511c603 fix(evals): preserve credential destination clauses
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 13:27:02 -06:00
DottaandPaperclip 249b487a2d test(e2e): read paired emphasis before review outcome checks
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 12:29:25 -06:00
Dotta 4713327d22 fix(runner): clarify active task history discovery scope 2026-11-04 12:29:25 -06:00
DottaandPaperclip b111963000 fix(evals): use completed personal setup context for acknowledgements
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 11:57:25 -06:00
DottaandPaperclip 4a2c0e5dd9 test(evals): reject unsafe company optional-note placeholders
Keep the safe note label and reject a rendered request for credential text; preserve company oracle v4.

Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 11:24:20 -06:00
DottaandPaperclip 54643f501d test(evals): reject unsafe personal question placeholders
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 11:23:34 -06:00
DottaandPaperclip 78d36b0899 fix(evals): inspect rendered secret-question placeholder
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 11:23:34 -06:00
DottaandPaperclip 9b17aa2c98 test(evals): declare company addressed-human acknowledgement v4
Exercise canonical anyone audience and safe optional-note affordance while retaining enumerated answers and saved-reference proof.

Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 11:09:21 -06:00
DottaandPaperclip 0af00a9024 fix(evals): admit actionable personal secret acknowledgements
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 11:08:58 -06:00
DottaandPaperclip 639ddf3ad8 fix(evals): honor addressed human question authority
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 11:08:58 -06:00
DottaandPaperclip 5f5e2ca8f1 test(heartbeat): retain mutable recovery event fault fixture
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 09:41:25 -06:00
DottaandPaperclip 225145192e fix(heartbeat): preserve server-finished conversation replies
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 09:41:25 -06:00
DottaandPaperclip 2a74888341 fix(evals): distinguish unbound roles from different actors
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 09:32:18 -06:00
DottaandPaperclip 3ab4c99afa style(evals): keep credential exclusion rationale in review
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 09:28:01 -06:00
DottaandPaperclip e4d6f9bc8f style(evals): keep personal handoff rationale in review
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 09:27:51 -06:00
DottaandPaperclip 855b98e005 chore(evals): remove added company credential comment
Keep the company helper executable source unchanged.

Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 09:27:51 -06:00
DottaandPaperclip 57509a283b chore(evals): remove added shared credential comments
Preserve baseline comments and executable source while following the contribution requirements.

Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 09:27:51 -06:00
DottaandPaperclip f3d21b7456 style(e2e): keep tool review rationale in the PR description
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 09:27:51 -06:00
DottaandPaperclip 3159a9ccd5 style(evals): keep timeline rationale in review description
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 09:27:51 -06:00
DottaandPaperclip 2692048d16 fix(evals): distinguish timeline dates from explicit runners
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 09:27:51 -06:00
DottaandPaperclip 3a6ced46b0 fix(heartbeat): keep native chat compatibility fallback closed
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 09:27:51 -06:00
DottaandPaperclip f14e6a2a62 fix(heartbeat): select native conversation finals by authority
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 09:27:51 -06:00
DottaandPaperclip a9c7c53724 fix(evals): preserve timeline structure across soft wraps
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 09:15:25 -06:00
DottaandPaperclip 07b67a517b test(evals): version personal confirmation semantics explicitly
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 09:15:25 -06:00
DottaandPaperclip 2b19fc60ab fix(evals): distinguish own-value completion from permission answers
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 09:15:25 -06:00
DottaandPaperclip 1c92458033 fix(heartbeat): deliver final handoff after progress promises
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 09:15:25 -06:00
DottaandPaperclip 34d8e441a4 test(e2e): match the approved replay provider error contract
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 09:07:11 -06:00
DottaandPaperclip 5409a10bad fix(evals): retain quoted timeline bullets as milestones
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 09:07:11 -06:00
DottaandPaperclip 9e087f06dc test(evals): distinguish deployment credential purpose from recipients
Normalize only the bound deployment key and its display name before named-recipient checks; retain foreign consumer rejection and saved-binding gates.

Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 08:57:09 -06:00
DottaandPaperclip 1d1589ce89 test(evals): bind named company secret completion recipients
Port only the reviewed company selector and focused controls from the e757 candidate successor. Keep ordinary completion language and the existing saved-binding proof.

Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 08:50:19 -06:00
DottaandPaperclip e757e44737 test(evals): report remaining fixed private scanner reasons
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 08:24:49 -06:00
DottaandPaperclip 5d91105991 fix(evals): inspect all rendered credential question instructions
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 08:24:49 -06:00
DottaandPaperclip 6b38cf596c fix(evals): preserve immediate timeline pronoun attribution
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 08:22:06 -06:00
DottaandPaperclip 1d2d8f732f fix(evals): accept concise bound personal setup confirmations
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 08:17:02 -06:00
DottaandPaperclip 3ad264635a test(evals): accept completed company setup acknowledgements
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 08:13:19 -06:00
DottaandPaperclip 8b5f90326d test(teams): provide company lookup in catalog service fixture
Match the saved-resource link lookup already present in the candidate. Preserve the missing-company rejection before import.

Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 08:12:26 -06:00
DottaandPaperclip c153d14f12 fix(evals): bound timeline pronouns to their immediate actor
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 08:12:26 -06:00
DottaandPaperclip d4ca3a80a0 test(evals): opt company secret proof into human answer accounting
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 08:12:26 -06:00
DottaandPaperclip 912a1e1df1 test(evals): accept bound company-secret question handoffs
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 08:12:26 -06:00
DottaandPaperclip 5e8e36958b fix(evals): scope question accounting to canonical credential cases
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 08:12:26 -06:00
DottaandPaperclip 244facf29d fix(evals): verify secure answers from resolver audit and delivery
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 08:12:26 -06:00
DottaandPaperclip 3fcf06adc8 test(evals): account for authenticated secure-entry question answers
Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-11-04 08:12:26 -06:00