* fix(install): skip Gemini local commands/gsd when global GSD present (#3037)
Reporter showed that running `npx get-shit-done-cc --gemini --global`
followed by `--gemini --local` in a project creates the same 65 GSD
command files in both Gemini scopes:
- ~/.gemini/commands/gsd/ (user scope)
- <project>/.gemini/commands/gsd/ (workspace scope)
Gemini conflict-detects by command name across scopes and renames every
overlapping /gsd:* command to /workspace.gsd:* and /user.gsd:*, breaking
the documented /gsd:* namespace.
Fix: in bin/install.js, when handling --gemini --local, detect whether
~/.gemini/commands/gsd/ already exists with managed-shape content. If
so, skip the local copy and print a clear three-line warning explaining
the conflict avoidance. The user-scope install already provides the
same /gsd:* commands in this project; the local copy adds zero value.
Sibling fixes (test isolation):
- tests/install-minimal-all-runtimes.test.cjs: pass HOME/USERPROFILE
through the spawned installer's env so the developer's real
~/.gemini/commands/gsd/ doesn't trigger the new skip path during
test runs that want to assert the local-install populates
commands/gsd/.
- tests/gemini-namespacing.test.cjs: the "Gemini Install (Behavioral)"
describe block now creates an isolated tmpHome and points
process.env.HOME at it before calling install(false, 'gemini'),
with proper restore in afterEach.
Test:
- tests/bug-3037-gemini-duplicate-commands.test.cjs — 4 structural
tests:
1. global install populates HOME/.gemini/commands/gsd
2. local install AFTER global skips the local copy
3. local install with NO existing global still populates locally
(no-regression)
4. local install when HOME has .gemini/ but no GSD-managed
commands/gsd/ still populates locally (non-GSD-Gemini-user
no-regression)
6909/6909 full suite pass. Lints clean.
Closes#3037
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
* fix: address CR feedback on PR #3041 — narrower detection + USERPROFILE restore
CR findings:
1. **bin/install.js (Major)** — userScopeHasGsd used
`fs.readdirSync(homeGeminiGsd).length > 0` which would skip the
local install for any non-empty directory, including a user who
hand-dropped a single override at ~/.gemini/commands/gsd/<thing>
.toml without ever running --gemini --global. Narrowed the
detection to require at least 3 canonical GSD command files
(help.toml, progress.toml, new-project.toml) — a marker that
ships in every GSD Gemini install (minimal mode included) and is
structurally impossible to produce by accident.
2. **tests/bug-3037-...:59 (Minor)** — beforeEach overwrites
process.env.USERPROFILE but afterEach only restores HOME, leaking
the temp home into later tests on Windows or any code path that
reads USERPROFILE. Added save/restore symmetric with HOME.
Plus added a 5th regression test covering the narrowed detection:
"local install when HOME has hand-dropped overrides UNDER commands/gsd/
(but no full GSD) still populates locally" — directly exercises the
edge case CR identified.
5/5 targeted tests pass. 6910/6910 full suite pass.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
Closes#2876 follow-up — CI on main fails because the punctuation test
in tests/gemini-namespacing.test.cjs hardcoded `/gsd-scan` as a known
command, but #2824 (consolidate 86 → 59 skills) removed scan.md from
commands/gsd/. The roster now correctly returns "scan is unknown, leave
unchanged" — the conversion is right, the test fixture is stale.
Swap `scan` for `health` in the punctuation test. Both are bedrock
commands; the test still exercises the original intent (period vs
exclamation handling on adjacent slash commands).
Note added so the next consolidation reviewer knows the swap pattern.
`npm test`: 5936/5936 pass.
* fix(install): use colon namespace for Gemini slash commands and help reference
This fixes unexecutable command recommendations in Gemini CLI by correctly
namespacing slash commands (/gsd: instead of /gsd-) in all installed
artifacts (agents, commands, workflows).
- Implements a lazy command roster discovery to ensure 100% accurate
conversion and protect file paths, URLs, and agent names.
- Adds isolated behavioral and unit tests covering all boundary cases.
- Fixes hardcoded command strings in banners and help output.
Closes#2783
* fix(install): close roster gaps in Gemini /gsd- → /gsd: conversion (#2783)
Addresses adversarial review findings on PR #2768:
- Restore regex boundaries (lookbehind + extension lookahead). Roster-only
matching was insufficient: a URL like `https://example.com/gsd-plan-phase`
ends in a known command and would be incorrectly converted. Boundaries +
roster now agree before any conversion fires.
- Smarter trailing lookahead `(?!\.[a-z])` distinguishes file extensions
(`.cjs`, `.md`) from sentence-ending punctuation (`.` at end of input or
before whitespace), so `/gsd-help.` correctly converts.
- Fail loud on missing roster. `commands/gsd/` not found previously fell
through to an empty Set, silently no-op'ing every conversion — exactly the
bug this code exists to prevent. Now emits a one-shot console.warn (gated
on GSD_TEST_MODE) before returning the empty set.
- Drop unnecessary `i` flag — GSD commands are always lowercase; matching
uppercase tokens against a lowercase roster always misses anyway.
- Export `_resetGsdCommandRoster` for test isolation against the module-level
cache.
Test additions pin the actual safety property of the roster check by using
KNOWN command names embedded in URLs and sub-paths — the cases the prior
tests didn't reach because they used `gsd-tools` (not in roster). Added a
roster-load assertion that fails loudly if the empty-Set fallback path
silently neutralises conversions.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
* fix(install): centralize <sub> stripping and add structural test assertions
CodeRabbit findings on the prior commit:
- (actionable) Centralizing the Gemini conversion through
convertClaudeToGeminiMarkdown dropped the stripSubTags() call that the
inline command path used to make before TOML conversion. Move stripSubTags
inside convertClaudeToGeminiMarkdown so command/agent/non-command Gemini
outputs all have <sub> consistently stripped. Remove the now-redundant
stripSubTags call in convertClaudeToGeminiAgent (single source of truth).
- (nitpick) Replace `.includes()` checks in the TOML test with structured
parsing — JSON-decode each TOML value and assert on parsed fields, per
the project's "tests parse, never grep" convention.
- (nitpick) Strengthen the install behavioral test to read a real installed
artifact (.gemini/commands/gsd/plan-phase.toml), parse it, and assert the
prompt body actually contains a /gsd: reference and no unconverted
/gsd-plan-phase. A directory-only check would have passed even if every
conversion silently no-op'd.
- Add a regression test that <sub> tags are stripped through the
convertClaudeToGeminiMarkdown pipeline.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
---------
Co-authored-by: Tom Boucher <trekkie@nomorestars.com>
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
