Skip to content

skills: tri-pipeline lists only commands that exist; analog-emulation gets the toolchain map and the loop tick protocol (Closes #10499) - #10500

Merged
gHashTag merged 1 commit into
masterfrom
claude/skills-toolchain-10499
Oct 18, 2026
Merged

gHashTag merged 1 commit into
masterfrom
claude/skills-toolchain-10499

Conversation

@gHashTag

Copy link
Copy Markdown
Owner

Closes #10499

tri-pipeline

Rewritten from --help of target/release/tri and t27c 0.5.2 (2026-10-18). Removed: tri verify (exists nowhere), tri experience query (only save exists). Corrected: scripts/tri test is the whole-repo t27c suite; the one-spec runner is target/release/tri test <spec>; t27c test <spec> only lists tests. Added: t27b verilog (refuses signed types), tri t27o, tri topic, tri loop claim, tri pr ready|landed, tri seals, tri hooks, the MCP servers and what each really checks. The root-level tri-pipeline.md becomes a pointer, so there is one list.

Commands named, each seen in a --help output on this date: t27c check, test-report, test, seal (--save, --verify), gen-verilog, gen-c, gen-rust, gen, fpga-build, fpga-flash; tri test, gen, t27o, topic, loop claim, pr ready, pr landed, night, red, ci, gates, merging, lanes, seals, hooks, experience save, loop-help (scripts/tri); t27b verilog, test (cli/t27b/src/main.rs usage).

analog-emulation

Delta-form step and proved stability (from #10487's design), unsigned arithmetic in hardware, the SPICE reader as a t27o front end (one converter, DRY), a toolchain table, and the six-step overnight loop tick: read the ledger, look for anomalies first, Opus review / Sonnet lanes, merge only green, write state, partner reports only on the owner's yes.

Related defects filed from the same survey: #10497 (tri test ignores vacuous passes), #10498 (no ax7203 in fpga-build / fpga-flash).

🤖 Generated with Claude Code

… gets the toolchain map and the loop tick protocol (Closes #10499)

tri-pipeline named `tri verify` (does not exist), `tri experience query`
(only `save` exists) and said `tri test` runs one spec (scripts/tri test
is the whole-repo suite). Rewritten from `--help` of target/release/tri
and t27c 0.5.2 on 2026-10-18; the root-level tri-pipeline.md becomes a
pointer so there is one list.

analog-emulation: delta-form step and proved stability, unsigned in
hardware (t27b verilog refuses signed types), the SPICE reader as a t27o
front end, a toolchain table, and the six-step overnight loop tick
(state, anomalies, Opus review / Sonnet lanes, merge green, write state,
reports only on the owner's yes).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@github-actions

Copy link
Copy Markdown
Contributor

PR Dashboard

Generated at: 2026-10-18 15:25:56 UTC

Summary

Status Count
Total Open PRs 50
PRs with Failing Checks 50
PRs with All Checks Green 0
READY 0
FAILING 50
PENDING 0
NO CHECKS YET 0

Seal Status

  • ⚠️ STALE -- sha256(compiler.rs)=fad12492937a != manifest seal=87e5cbd3ad94.
    The committed NMSE numbers were certified against an older compiler.rs.
    Run scripts/reseal-check.sh locally for the two-step reseal command (advisory; not a merge gate).

@gHashTag
gHashTag merged commit 3c24724 into master Oct 18, 2026
28 of 30 checks passed
gHashTag added a commit that referenced this pull request Oct 18, 2026
… mentions (Closes #10513) (#10514)

#10500 removed `tri verify` and `tri experience query` from
.claude/skills/tri-pipeline.md, so the equality ratchet read 138 against
the ceiling 140 and "Documented t27c subcommands exist" failed on master
and on every PR after it. One constant inside a gate tool already listed
in tools/policy/foreign-exceptions.txt (the PR #9277 pattern).
`python3 tools/check_documented_commands_exist.py` now prints "ok".

Co-authored-by: Claude <claude@anthropic.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

2 participants