Repository navigation
Specify contributor key management and attribution boundaries - #5473
Merged
Merged
Conversation
Contributor
|
📓 NotebookLM Notebook linked to this PR
This notebook contains session context, decisions, and artifacts for this work. |
Contributor
Refs #5472 phi^2 + 1/phi^2 = 3 | TRINITY
Contributor
|
📓 NotebookLM Notebook linked to this PR
This notebook contains session context, decisions, and artifacts for this work. |
Contributor
PR DashboardGenerated at: 2026-10-01 22:11:19 UTC
Summary
Seal Status
|
dmitrii-f-t27
marked this pull request as ready for review
October 1, 2026 22:12
This was referenced Oct 1, 2026
Contributor
|
📓 NotebookLM Notebook linked to this PR
This notebook contains session context, decisions, and artifacts for this work. |
Contributor
PR DashboardGenerated at: 2026-10-02 00:12:56 UTC
Summary
Seal Status
|
Contributor
PR DashboardGenerated at: 2026-10-02 00:25:23 UTC
Summary
Seal Status
|
Contributor
|
📓 NotebookLM Notebook linked to this PR
This notebook contains session context, decisions, and artifacts for this work. |
dmitrii-f-t27
force-pushed
the
codex/hive-contributor-keys
branch
from
October 2, 2026 00:39
1a5e556 to
2bc83bf
Compare
Contributor
PR DashboardGenerated at: 2026-10-02 00:39:25 UTC
Summary
Seal Status
|
Contributor
|
📓 NotebookLM Notebook linked to this PR
This notebook contains session context, decisions, and artifacts for this work. |
…ive-contributor-keys
…#5472) At the reviewer's request (changes requested on 9eab05d) the PR is split. Restored to master: the 80 placeholder tests of igla/coder/pipeline and gf16_matmul_top with the ratchets that follow them (now #5613), the two seals of specs that had none (verilog_ternary_mac_top, vivado_gf16_matmul_top), the akashic-log claim lines, and the corpus re-take and test-block pin, which move with every port PR and are redone right before merge. What remains: the two contributor-key specs, their controls, their two seals and the docs/now entry. Both specs pass 17 tests and `seal --verify`. Refs #5472 Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Contributor
|
📓 NotebookLM Notebook linked to this PR
This notebook contains session context, decisions, and artifacts for this work. |
Contributor
PR DashboardGenerated at: 2026-10-02 16:58:10 UTC
Summary
Seal Status
|
dmitrii-f-t27
added a commit
that referenced
this pull request
Oct 2, 2026
…efs #5472) (#5613) specs/igla/coder/pipeline.t27 carried 80 tests whose body is only `assert true` / `{ /* verify baseline */ }`; gf16_matmul_top.t27 two comment-only tests. They run nothing. Removed, pipeline resealed with the t27c built from this tree, and the ratchets follow: corpus ledger drops gf16_matmul_top (134/134, CLEAN), assertionless baseline 3905 -> 3761, published test blocks 14330 -> 14314 (master measured 14394). Split out of #5473 at the reviewer's request. Refs #5472 Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
Contributor
Author
|
Scope split as requested in the review of What this PR carries now (7 files against master Where the rest went
CI on The merge is blocked only by the change request on |
This was referenced Oct 2, 2026
gHashTag
added
the
bee-reviewed
Owner
|
Reviewer bee (triage 2026-10-03): re-review at head
|
![t27-bees[bot]](https://avatars.githubusercontent.com/in/5166765?s=60&v=4){kind=small}
dmitrii-f-t27
added a commit
that referenced
this pull request
Oct 10, 2026
The standalone matcher in tools/check_documented_commands_exist.py read every hyphenated tri- word as a missing scripts/tri-<name>: 1004 mentions under 113 names on master, all of them skill names (tri-pipeline), agent ids (tri-doctor), sibling repositories (../tri-net/src/lib.rs) or the adjective tri-valued. None is a command, and the job has failed on every master run. The repair is the one from #5473: a standalone name counts only in command position (line start, optionally after "$ " or "> ", or after a backtick, followed by an option or the end of the line), and the scan reads tracked files from git ls-files instead of walking the working tree. scripts/tri-* paths are matched as before. On master d5f2215: the check passes (11 sibling references, 5 excused as declared, 141 dead tri mentions at the recorded ceiling) and --self-check passes, including 11 new sibling controls. Refs #5497
dmitrii-f-t27
added a commit
that referenced
this pull request
Oct 10, 2026
check_specs_generate.py failed on master: 15 specs under specs/port/ do not generate with any backend, each merged while this step was already red. Making them parse is not a repair -- on #5473 seven were made to parse and their generated C and Zig still had dozens of errors per file -- so their ports are to be redone, tracked in #5549 with the error and the PR that added each one. Until then they join tools/specs_generate_baseline.txt by hand, with the compiler's own first line, which is how this ledger is meant to grow (--update-baseline refuses to grow it). 14 -> 29 entries. The gate passes on 769f325 (1212 specs, 1183 generate) and its self-check passes. Refs #5497
dmitrii-f-t27
added a commit
that referenced
this pull request
Oct 10, 2026
Eleven groups were in no ledger on master 756bcff. Scanner. tools/dupe_scan.py now reads the lexical forms t27c has: a declaration ending in ';' no longer borrows the next function's body (digest_for x8 and githubCiDeps_checkRuns_impl x5 were that), comments and string literals are masked, ';' at column 1 and '#' are line comments, and a single quote ends at the newline as the lexer ends it. The prototype, comment and literal repair comes from #5473. The quote and ';' rules are new: an apostrophe in "; the Queen's ..." opened a literal that hid three bodies in specs/queen/views.t27 and specs/tools/catalog.t27. Eleven self-test shapes pass; the #5473 lexer fails the two new ones. forwardPass x3: one empty-result literal in specs/hslm/forward_pass.t27 is shared (from #5473). Both seal files naming the spec are resealed with the t27c built from this tree; all four gen hashes equal the #5473 seals. rng_int_range x2: the gen_fuzz port had two 'return undefined;' stubs. They are now a xorshift64 generator (shifts and xors only: t27c emits u64 '+' and '*' without wrapping in Zig and Rust), and the port has five tests that can fail. Zig: 8/8 pass; five mutants are each caught. Seven copies across two files are ledgered, as #4497 did, because 'use a::b;' still generates no import (#4610): bytes_eq, session_status_str, led_config, validate_led_config, startup_config, update_oscillator_chain, shift_ir. Three rows whose groups no longer exist are dropped: generate_all, next, is_orphan. Refs #5497
dmitrii-f-t27
added a commit
that referenced
this pull request
Oct 10, 2026
…efs #5472) specs/igla/coder/pipeline.t27 carried 80 tests whose body is only `assert true` / `{ /* verify baseline */ }`; gf16_matmul_top.t27 two comment-only tests. They run nothing. Removed, pipeline resealed with the t27c built from this tree, and the ratchets follow: corpus ledger drops gf16_matmul_top (134/134, CLEAN), assertionless baseline 3905 -> 3761, published test blocks 14330 -> 14314 (master measured 14394). Split out of #5473 at the reviewer's request. Refs #5472
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The game account needs a trusted boundary for contributor credentials and attribution of existing Queen work. Define render and Queen contracts for verified person sessions, request limits, exact action routes, stable key IDs, immutable ownership and bounded probes. The production hosts already consume these generated constants.
This update also repairs the first existing CI blockers found while preparing the canonical contracts for landing:
8edc1c43; the scoped follow-up below reduces this to eight.assert truefrom three historical specs. Non-test source is unchanged. Lower the assertionless ledger from 4049 to 3761, including the already-resolved cordic entry; the guard and negative controls pass. Regenerate and verify only the three changed specs' seals. The counter spec's five C tests pass.Validation on M1 at
8edc1c43(merged base92d4b247):t27candtripassed. All 17 contributor-contract Zig tests and both feature seals passed again.GitHub at
8edc1c43confirms Documented Commands and all three required contexts (validate,check-linked-issue,parse-ratchet) as SUCCESS. The assertionless step in Corpus Ratchet also passes. Exact-head generation findings are byte-identical to the verified base (15); the full seal comparison adds no stale finding, changes none of the other stale records, and removes the repaired pipeline seal (591 stale, 655 current).Full repository CI remains blocked. The remaining work includes eight duplicate-body groups, the preexisting generation failures, stale seals, 15 unjudged type conflicts (and three resolved ledger entries), ring-096 signature drift, nine outdated published corpus measurements, and the corpus suite's existing expectation differences. Later stages were executed locally to expose failures hidden behind the first red step. Local FPGA smoke output is synthetic/dry-run evidence, not a physical-board result. The full local suite was run before the exploratory syntax edits were reverted and before the final three scoped seals were regenerated; it is diagnostic evidence, not an exact-head all-green receipt.
No failing context is reported as green. No duplicate ceiling, generation baseline, branch protection or merge guard was raised or bypassed. The PR remains open because repository policy treats failing tests as blocking.
Final scoped follow-up at
1a5e5564: factor the three existing emptyForwardOutputinitializations into one helper. Public signatures and placeholder behavior are unchanged; this does not implement inference. The duplicate-body guard now reports eight groups, and all nine scanner controls pass. An isolated harness compiles the generated C types/functions verbatim, with-Dnull=NULLfor the existing backend spelling, and exercises 14 calls both before and after; a non-null-loss mutant fails. This bounded check does not certify the full incomplete module. Both seal records naming this spec were regenerated and verified. The assertionless count stays 3761.The owner explicitly requested one small final block, preservation of the remaining usage allowance, and a Claude Code handoff. Leave this PR OPEN. CI for the latest push has started; no all-green claim is made.
Closes #5472. Related: #3141 and gHashTag/999-multibots-telegraf#3251.
phi^2 + 1/phi^2 = 3 | TRINITY