Triage of foundry-feedback.ledger.yml for run draft-manuscript-galaxy, produced by
report-foundry-run-feedback. All upstream reports target galaxyproject/foundry, including
those whose root cause is a related project.
<run>/foundry-feedback.ledger.yml (25 entries, all status: open)main checked at the time of triage; tracker searched across open and closed issues.| Field | Value |
|---|---|
run.pipeline |
paper-to-galaxy |
run.run_slug |
draft-manuscript-galaxy |
run.status |
complete |
All 12 top-level phases are status: done and feedback_checked: true:
| # | Skill | Status | Checked |
|---|---|---|---|
| 1 | summarize-paper |
done | yes |
| 2 | freeform-summary-to-galaxy-interface |
done | yes |
| 3 | freeform-summary-to-galaxy-data-flow |
done | yes |
| 4 | compare-against-iwc-exemplar |
done | yes |
| 5 | freeform-summary-to-galaxy-template |
done | yes |
| 6 | advance-galaxy-draft-step (loop, 9 iterations) |
done | yes |
| 7 | branch: test-data resolution | done | yes |
| 8 | freeform-summary-to-galaxy-test-plan |
done | yes |
| 9 | implement-galaxy-workflow-test |
done | yes |
| 10 | validate-galaxy-workflow |
done | yes |
| 11 | run-workflow-test |
done | yes |
| 12 | debug-galaxy-workflow-output |
done | yes |
Coverage verdict: complete and fully reviewed. Every phase reported a feedback outcome, so the ledger’s contents are evidence of what the run actually found rather than of what it failed to look for. No phase is unchecked and none went unreported.
Four entries were raised outside the 12 pipeline phases — three by author-galaxy-tool-wrapper
and freeform-summary-to-galaxy-template during the post-pipeline deploy step, and one
(ga-download-normalizes-tool-id-producing-workflows-that-save-but-cannot-invoke) by
raised_by: manual-deploy, which is not a Mold and carries observed_in: null. See §3.
status: open with issue: null. Nothing was previously filed, so no
entry needed its prior disposition honored.observed_in block.
ga-download-normalizes-tool-id-producing-workflows-that-save-but-cannot-invoke has
observed_in: null and raised_by: manual-deploy, a non-Mold actor. Its subject.content_hash
is also null. Per the protocol, missing fields were not inferred; the draft for it states the
Foundry-side locator without an observed hash.subject.content_hash (paper-to-test-data-no-scoping-or-toolshed-fixture-guidance,
run-workflow-test-no-mechanism-to-install-authored-galaxyusertool-udts). For both, the
observed_in.mold.content_hash matches current main for the same path, so the subject content
was still verifiable; the drafts cite the observed-in hash and say which one it is.content_hash: unavailable-in-cast
(run-workflow-test-no-mechanism-to-install-authored-galaxyusertool-udts, run-workflow-test
revision 6). A later entry in the same run observed revision 7 of the same Mold at
259675e41c65…, which matches current main.major, 12 minor; 11 defect, 14 gap.author-galaxy-tool-wrapper-no-check-of-wrapped-scripts-own-declared-input-ordering-assumption
names a private third-party repository as the home of the wrapped script. That name is
removed from the upstream draft and replaced with “a private repository outside the Foundry”.
The observation survives redaction: the Foundry-side gap is in the authoring process, not in the
script, and is fully statable without naming the repository.galaxy.key file is not referenced by any entry and its contents appear nowhere.Clustered by subject.locator, per protocol. 25 entries → 25 clusters. Nothing merged.
Three locator collisions were examined and deliberately not merged, because each pair requests a distinct correction:
content/research/galaxy-workflow-invocation-failure-reference/index.md carries two entries.
galaxy-workflow-invocation-check-rejects-previously-installed-tool-as-not-installed asks for a
stale-toolbox-read diagnostic tell; ga-download-normalizes-tool-id-producing-workflows-that-save-but-cannot-invoke
reports a deterministic unversioned-tool_id defect and states in its own text that it is
distinct from the first (no timing component, reproduces on demand).jmchilton/galaxy-tool-util-ts, but at four different components
(ToolShed-fetch decoder, tool-search tokenizer, draft-validate --json emission,
validate --connections step builder) with four unrelated fixes.Every Foundry-subject locator was resolved against galaxyproject/foundry@main and its content
hash compared to the ledger’s.
| Locator | Ledger hash | main | Result |
|---|---|---|---|
content/patterns/galaxy-collection-patterns.md |
9840d686228a |
9840d686228a |
unchanged |
content/research/gxformat2-schema/index.md |
4fe42382f996 |
4fe42382f996 |
unchanged |
content/molds/implement-galaxy-tool-step/index.md |
197be2d9b277 |
197be2d9b277 |
unchanged |
content/research/galaxy-user-tool-authoring/index.md |
365fad455369 |
365fad455369 |
unchanged |
content/molds/paper-to-test-data/index.md |
(none; observed-in 863d7b503a77) |
863d7b503a77 |
unchanged |
content/molds/freeform-summary-to-galaxy-test-plan/index.md |
9c1d56e625a5 |
9c1d56e625a5 |
unchanged |
content/research/iwc-test-data-conventions/index.md |
1921e9397034 |
1921e9397034 |
unchanged |
content/molds/run-workflow-test/index.md |
(none; observed-in rev 7 259675e41c65) |
259675e41c65 |
unchanged |
content/research/galaxy-workflow-invocation-failure-reference/index.md |
e8115df6ddca |
e8115df6ddca |
unchanged |
content/molds/author-galaxy-tool-wrapper/index.md |
cced8f068cd1 |
cced8f068cd1 |
unchanged |
content/molds/advance-galaxy-draft-step/index.md |
c92d452f69a7 (rev 4) |
ae19f7e5f5ac (rev 7) |
changed — inspected |
The one changed locator was inspected rather than assumed. advance-galaxy-draft-step advanced
from revision 4 to revision 7 (revised: 2026-09-17). Its current text contains no occurrence of
sample_sheet, sample sheet, param_value_from_file, or any scalar-parameter-wiring rule, so the
correction requested by lexicmap-search-sample-sheet-column-to-scalar-port-defect is not
present on main. The observation is not marked fixed; its draft records both the observed hash
(rev 4) and the current successor hash (rev 7), so a maintainer reviews the live text.
The two package://@galaxy-foundry/gxwf-foundry#… schema subjects are package exports, not
repository paths; their ledger hashes are carried into the drafts as observed identity without a
path-based main comparison.
Related-project subjects (11 entries) have no Foundry locator and no hash. Per protocol, the main check was skipped for them rather than invented.
No observation was marked fixed. Nothing in current Foundry source contains the requested correction.
Searched galaxyproject/foundry across open and closed issues by canonical locator, subject label,
and correction. Three matches found; all three are open, so each becomes a comment draft rather
than a second issue.
| Cluster | Match | Disposition |
|---|---|---|
galaxy-workflow-import-enforces-undocumented-per-field-length-limit-on-step-doc |
#564 (open) — same defect: step doc: length crashes live Galaxy import, no static check catches it |
comment-draft |
gxwf-draft-validate-json-flag-emits-non-json-diagnostics-on-stdout |
#548 (open) — explicitly reports this same routing half of upstream galaxy-tool-util-ts#166 as still present |
comment-draft |
galaxy-user-tool-authoring-missing-element-identifier-expression |
#468 (open) — same authoring-reference surface, same expression-form table that the issue already faults | comment-draft |
On #564 the run contributes a materially different measurement and a conflicting threshold: #564 bisected the limit to between 3900 and 3999 characters; this run bisected it to exactly 2048 on usegalaxy.org 26.1.2.dev0. Both cannot describe one fixed constant, so the comment draft raises that discrepancy rather than merely seconding the report.
Near misses examined and rejected as duplicates, cross-referenced in the drafts instead:
.yml/.myst companions. Mechanically adjacent to the
pattern-MOC cluster but a different packaging path (sidecar files vs. wiki-linked sibling pages).
The closed fix does not cover it.tool-search returns duplicate rows. Same command as the repo-slug cluster,
different defect (dedup vs. tokenizer scoring). Not resolved by that fix.column_definitions. Import-side; the
test-data cluster is about the rows: job-input shape. Distinct.#566 (“Needs better guidance for importing to a real galaxy instance”) deserves its own note. It
is an open, unanswered request for exactly the guidance five of this run’s deploy-time clusters
produced empirically (frame-comment position/size, /api/unprivileged_tools registration,
tool_uuid injection, exact_tools/allow_missing_tools, versioned tool_id). Those five remain
separate issue drafts because they are five distinct corrections against four different components —
but a sixth draft is included that answers #566 directly and links them, so the request is closed out
rather than left standing beside five issues that happen to contain its answer.
| # | Cluster (entry id) | Sev | Disposition |
|---|---|---|---|
| 1 | collection-pattern-mocs-ship-without-referenced-pages |
minor | issue-draft |
| 2 | gxformat2-schema-missing-nested-in-key-convention |
major | issue-draft |
| 3 | tool-util-cli-toolshed-fetch-rejects-real-filtered-list-collection-output |
major | issue-draft |
| 4 | implement-galaxy-tool-step-udt-binding-undocumented |
minor | issue-draft |
| 5 | galaxy-user-tool-authoring-missing-element-identifier-expression |
minor | comment-draft → #468 |
| 6 | gxwf-tool-search-underscore-repo-slug-query-zero-hits |
minor | issue-draft |
| 7 | gxwf-draft-validate-json-flag-emits-non-json-diagnostics-on-stdout |
minor | comment-draft → #548 |
| 8 | paper-to-test-data-no-scoping-or-toolshed-fixture-guidance |
minor | issue-draft |
| 9 | galaxy-workflow-test-plan-schema-fixture-required-for-scalar-typed-params |
minor | issue-draft |
| 10 | freeform-summary-to-galaxy-test-plan-silent-on-concrete-draft-and-test-data-refs-inputs |
minor | issue-draft |
| 11 | tests-format-job-schema-has-no-path-to-an-intermediate-step-input-port |
major | issue-draft |
| 12 | tests-format-schema-silent-on-sample-sheet-column-definitions-shape |
minor | issue-draft |
| 13 | galaxy-workflow-frame-comments-missing-position-size-rejected-by-real-galaxy-import |
major | issue-draft |
| 14 | gxwf-validate-connections-flag-crashes-uncaught-on-format2-dict-shaped-step-in |
major | issue-draft |
| 15 | run-workflow-test-no-mechanism-to-install-authored-galaxyusertool-udts |
major | issue-draft |
| 16 | galaxy-tool-util-replacement-collection-requires-rows-key-unconditionally |
major | issue-draft |
| 17 | galaxy-workflow-invocation-check-rejects-previously-installed-tool-as-not-installed |
major | issue-draft |
| 18 | galaxy-udt-registration-real-endpoint-is-unprivileged-tools-not-dynamic-tools |
major | issue-draft |
| 19 | galaxy-workflow-import-enforces-undocumented-per-field-length-limit-on-step-doc |
major | comment-draft → #564 |
| 20 | galaxy-unprivileged-tools-lint-crashes-on-integer-input-default-value-zero |
minor | issue-draft |
| 21 | gxformat2-step-schema-does-not-model-tool-uuid-for-dynamic-tool-resolution |
major | issue-draft |
| 22 | lexicmap-search-sample-sheet-column-to-scalar-port-defect |
major | issue-draft (hash moved rev 4→7) |
| 23 | galaxy-workflow-update-exact-tools-defaults-true-and-discards-per-step-error-detail |
major | issue-draft |
| 24 | author-galaxy-tool-wrapper-no-check-of-wrapped-scripts-own-declared-input-ordering-assumption |
major | issue-draft (redacted) |
| 25 | ga-download-normalizes-tool-id-producing-workflows-that-save-but-cannot-invoke |
major | issue-draft |
| — | deploy-guidance synthesis answering #566 | — | comment-draft → #566 |
Totals: 22 new-issue drafts, 4 comment drafts, 0 fixed, 0 duplicate-closed, 0 local-only, 0 wontfix.
A complete, fully-reviewed run whose feedback is unusually deployment-heavy: 11 of 25 observations
were only reachable by pushing the finished workflow at a real Galaxy instance, and 8 of those are
Galaxy-core or harness-CLI defects rather than Foundry content gaps. The pipeline’s own static gates
(gxwf validate, gxwf draft-validate --concrete) passed clean on a workflow that could not be
imported, could not be invoked once imported, and contained a wiring defect
(sample_sheet column → scalar tool port) that no local check can currently catch. That gap between
“passes every Foundry gate” and “runs on real Galaxy” is the through-line, and it is the same gap
open issue #566 is asking about.
Nothing in the ledger has been fixed upstream since the run. One Mold changed on main but not in the respect this run faulted. Three observations land on existing open issues and are drafted as comments; the rest are new.
Per protocol, no remote mutation has been performed. foundry-issue-drafts.md holds the bodies;
filing awaits explicit confirmation of the exact set, after which the corresponding ledger entries
are updated to filed with their issue URLs.