docs(roadmap/backlog): priority install-friction plugins and bundled-plugin policy

[tonythethompson]

Add an explicit install-friction priority queue to the plugin roadmap (compress, terminal_qr/qr_maker, nugins endecode, units, hashes, kdl, plotters/plot, dbus) and a core/bundled plugins policy.

Remove FMotalleb/nu_plugin_clipboard from P4 maintained forks because Nushell 0.111+ provides the clip native command.

Update backlog source data: correct repos for
u_plugin_endecode,
u_plugin_hashes, and
u_plugin_plotters; set
u_plugin_formats to BUILTIN; add priority notes and version data for the listed candidates.

Generated with Devin

[github-actions]

Review of PR #102

Overall: Documentation-only changes, internally consistent. JSON is valid and properly formatted; check_repo_consistency.py passes; the roadmap cross-link to consolidated-multi-repo-roadmap is preserved. The repo corrections (endecode → kaathewisegit/nugins, hashes → ArmoredPony/nu_plugin_hashes, plotters → cptpiepmatz/nu-jupyter-kernel) and the BUILTIN marking for nu_plugin_formats are sound, and every priority list entry has a matching backlog record.

One substantive concern — the clip claim is imprecise. I checked the Nushell 0.111.0 release blog. The native clip copy / clip paste commands exist, but they are experimental and gated behind --experimental-options=native-clip (PR #17572 / #17664) — not a default-on built-in. The roadmap states it "supersedes FMotalleb/nu_plugin_clipboard" and the backlog note calls the plugin "not a numan candidate," and the PR drops it from the P4 fork queue entirely. For users who don't opt into the experimental flag, the plugin is still the practical path until native-clip stabilizes to default-on. Consider softening to "supersedes once native-clip is stable / default" rather than a hard removal, or note the experimental gate explicitly in the policy section.

Minor nits:

• docs/roadmap.md:112 lists Euphrasiologist/nu_plugin_plot as a fallback plotting option, but that plugin is already shipped ([x] at line 172). The phrasing "if it remains the better maintained option" reads as a future candidate when it's already in the registry — clarify it's an already-available alternative.
• devyn/nu_plugin_dbus backlog note now says "evaluate numan-maintained fork," but devyn is the upstream owner, not a fork target — "upstream bump" alone fits better.

Neither nit blocks merge. The clip wording is the one I'd want addressed before merge, since it's the load-bearing justification for dropping a maintained-fork candidate.

opencode session  |  github run

[tonythethompson]

/oc fix

[github-actions]

fatal: could not read Username for 'https://github.com': No such device or address

opencode session  |  github run

[tonythethompson]

/oc fix your opencode.yml config so you stop quietly failing at the tasks your assigned

[cubic-dev-ai]

2 issues found across 2 files

Confidence score: 4/5

• In docs/backlog.json, the ELIGIBLE entry has an empty backfill_targets, so releases 0.113 and 0.112 will be omitted from the backfill queue; populate the targets or change the status.
• In docs/backlog.json, the native-clip note excludes a plugin that the roadmap identifies as the practical path while default-on remains gated, which could make backlog guidance misleading; update the timing or clarify the P4-only scope.

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="docs/backlog.json">

<violation number="1" location="docs/backlog.json:31">
P2: While `native-clip` is still gated, this note excludes the plugin from consideration even though the roadmap says it remains the practical path until default-on. Change the timing, or clarify that only the P4 maintained-fork queue is excluded.</violation>

<violation number="2" location="docs/backlog.json:790">
P2: Because this entry is `ELIGIBLE` and lists unpromoted 0.113/0.112 releases, an empty `backfill_targets` hides both backfill opportunities from the queue. Populate it with `["0.113", "0.112"]`, or change the status if these releases are intentionally ineligible.</violation>
</file>

<sub>Reply with feedback, questions, or to request a fix.

Fix all with Cursor | Re-trigger cubic</sub>

[github-actions]

fatal: could not read Username for 'https://github.com': No such device or address

opencode session  |  github run

[github-actions]

fatal: could not read Username for 'https://github.com': No such device or address

opencode session  |  github run

[tonythethompson]

/oc fix

[github-actions]

fatal: could not read Username for 'https://github.com': No such device or address

opencode session  |  github run

[github-actions]

Both cubic review issues are resolved:

1. docs/backlog.json:31 (native-clip note) — already addressed in bb673a6; the note now states native-clip is gated and the plugin "remains the practical path" until default-on.
2. docs/backlog.json:790 (empty backfill_targets) — fixed: the ELIGIBLE nu_plugin_endecode entry lists unpromoted 0.113.1/0.112.0 releases, so I populated backfill_targets with ["0.113", "0.112"], matching the schema field definition and the existing nu_plugin_hmac/nu_plugin_bigquery pattern.

JSON validates and check_repo_consistency.py passes.

opencode session  |  github run