Skip to content

lab: seal_new input saves first seals with the lab t27c - #112

Merged
gHashTag merged 1 commit into
masterfrom
lane/lab-seal-new
Nov 9, 2026
Merged

gHashTag merged 1 commit into
masterfrom
lane/lab-seal-new

Conversation

@gHashTag

@gHashTag gHashTag commented Nov 9, 2026 •

Copy link
Copy Markdown
Collaborator

Refs Trinity-S-AI-DNA/t27#12871

The lab seal leg runs only the drift repair, so a spec that has no seal yet never gets its first one. New t27b specs (ifconv_plan, middle_end, soon flow_plan) sit unsealed.

This adds one dispatch input, seal_new (space-separated spec paths, default empty = off). For each path a new step saves a seal with the lab's own t27c from the t27 checkout root (so spec_path is repo-relative, no home path), then copies new or changed seals into runs/<sha>.seals/ and appends them to CHANGED.txt, the same way the drift leg does. The existing seal step is untouched (it is a byte copy of SEAL_STEP_YAML in t27 specs/t27os/lab_seal.t27). The spec list is passed through env, not interpolated into the script.

Foreign budget: lab.yml +26 / -26 (net zero, under 40). The deletions are comment lines condensed to the same meaning and one stale "re-registration touch" note; no step logic was removed.

Not run locally (lab only).

🤖 Generated with Claude Code

@gHashTag
gHashTag merged commit df525a5 into master Nov 9, 2026
2 checks passed
@gHashTag
gHashTag deleted the lane/lab-seal-new branch November 9, 2026 06:27
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant