[DanielNetzer]

Implements issue #292.

Molecule: mol-292-4273a4a07a89474afb667f81
Formula: forgecell-github-issue
Base: 33dfa8f3530aae2057b668a6295be2ec2d2dd6aa
Delivery: 124d67ac8035687342396203d674a41ef74a280f

Awaiting exact-commit CI and human review. Harness-reported checks are not independent CI evidence. No merge, deployment or issue closure is authorized.

[DanielNetzer]

Adversarial review (plugins slice 1a): no HIGH or MEDIUM findings, and no behaviour drift. I checked:

• readIssue → tracker.Get: argv, Dir, empty Stdin, the 20s timeout, the default cap, the error text, and the partial-Issue-plus-decode-error path.
• The Ticket↔Issue field copy.
• States defaults: zero Timeout and MaxBytes fall back to 20s and 4,000,000 bytes.
• The List default signatures.
• DisplayKey output: byte-identical at all three switched sites, while the shell-command sites stay on Number.
• That Issue stays comparable and keeps its JSON shape (Key is omitempty and never set), and the import direction.
• main_test.go TMPDIR handling.

LOWs, to handle in slice 2b:

1. delivery/publish.go:341 still writes Implements issue #%d. from Number, while the demo prBody now uses DisplayKey(). That's identical for GitHub, but the preview and the published body would diverge for a plugin tracker.
2. cli/queue.go:106 prints %s#%d from an intake.QueueIssue, which is not a molecule.Issue. Revisit it when the queue lists tracker keys.

Integration: I applied this branch to current main (with #295 and #296) and go vet ./... passes. These packages also pass: go test -timeout 20m ./internal/{tracker,molecule,intake,forgesync,ui,cli,demo,delivery}/....