mirror of
https://github.com/garrytan/gstack.git
synced 2026-11-22 16:11:34 +01:00
Codex picker prompts and docs name skills the way Codex lists them (GST-3196) (#4524)
On Codex, the suggested prompts for nine gstack skills and the README now say `$office-hours`, `$review` and so on, which load the skill; `$gstack-office-hours` loaded nothing. Why: Codex lists each skill by the `name:` in its SKILL.md, not the `gstack-` folder, and skill_prefix never changes that name on Codex. Nothing to do: the next `./setup --host codex` picks it up. Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
This commit is contained in:
1 parent
b0f865c00c
commit
b23e1b01a7
16 files changed
+52
-17
No files matched your search
+2
-2
@@ -167,7 +167,7 @@ after you approve); ask for `gstack-office-hours`, or list them with
|
||||
Codex skills in `.agents/skills` (the workspace tier wins), and workspace skills
|
||||
load only in trusted folders (`/trust`).
|
||||
|
||||
Outside reviews require the selected CLI to be installed and authenticated: Claude Code when using gstack in Codex, or Codex on other harnesses. External harnesses discover these commands as `/gstack-claude-code` and `/gstack-codex`; each harness omits its own wrapper. Explicit provider requests keep that provider. The existing `codex_reviews` setting controls automatic outside reviews where supported, regardless of the provider selected.
|
||||
Outside reviews require the selected CLI to be installed and authenticated: Claude Code when using gstack in Codex, or Codex on other harnesses. External harnesses discover these commands as `/gstack-claude-code` and `/gstack-codex` (Codex lists its one as `claude-code`: mention it as `$claude-code`); each harness omits its own wrapper. Explicit provider requests keep that provider. The existing `codex_reviews` setting controls automatic outside reviews where supported, regardless of the provider selected.
|
||||
|
||||
`/claude` has been renamed to `/claude-code`. Re-run `./setup --host <name>` to migrate managed installations, including other harnesses sharing the checkout. Setup preserves the previous installation if replacement generation or installation fails and prints repair instructions.
|
||||
|
||||
@@ -226,7 +226,7 @@ repair message and reports the outside review as unavailable. It never silently
|
||||
switches to its default. Nested Codex reviews also run with installed skills
|
||||
hidden (`-c skills.include_instructions=false`), so a review cannot turn into a
|
||||
whole nested skill run. Runtime model selection is separate from the setup-time
|
||||
behavioral profile above. Without the plan-review role, `/claude-code` (`gstack-claude-code`
|
||||
behavioral profile above. Without the plan-review role, `/claude-code` (`$claude-code`
|
||||
on Codex) preserves Claude's configured model. Set `GSTACK_CLAUDE_MODEL=<model>`
|
||||
or name a model in your request to override it for the invocation, including
|
||||
resumed consultations. See [eval defaults and overrides](../CONTRIBUTING.md#testing--evals)
|
||||
|
||||
@@ -95,7 +95,7 @@ The hook reads `destructive_guard` every time it runs, so `gstack-config set des
|
||||
|
||||
**Contracts on Codex** use the same Stop hook, with the same one block per skill run, records and fail-open rules; Codex sends it the same `stop_hook_active` and `last_assistant_message` keys and takes the same block answer. The transcript is Codex's rollout file (`~/.codex/sessions/<yyyy>/<mm>/<dd>/rollout-*.jsonl`), and the checker tells it from a Claude Code transcript by what the file holds (a rollout starts with a `session_meta` record), never by its path. On Codex:
|
||||
|
||||
- A **load** is Codex injecting a `gstack-*` skill into the conversation as a `<skill>` message (the app server's skill input does this), or the model reading one with a shell command such as `cat ~/.codex/skills/gstack-review/SKILL.md`. `codex exec '$gstack-review ...'` passes the `$` mention as plain text, so in `exec` the skill loads only when the model reads it.
|
||||
- A **load** is Codex injecting a `gstack-*` skill into the conversation as a `<skill>` message (the app server's skill input does this), or the model reading one with a shell command such as `cat ~/.codex/skills/gstack-review/SKILL.md`. `codex exec '$review ...'` also injects the skill as a `<skill>` message (codex-cli 0.170); a mention of the folder name instead of the listed name is plain text, so then the skill loads only when the model reads it.
|
||||
- A **dispatch** is a `spawn_agent` call that started an agent. One that failed (its result has no `agent_id`) doesn't count.
|
||||
- **The contract** comes from gstack's own copy of the skill (`<gstack>/review/SKILL.md`), because Codex's installed `SKILL.md` keeps only its name and description.
|
||||
- **/review's `specialists-dispatched` item is skipped**, because Codex doesn't run /review's specialist reviewers yet. The adversarial pass and the report are checked as on Claude Code, and the block text names `spawn_agent` and `wait_agent` where Claude Code's names an Agent dispatch.
|
||||
|
||||
+2
-2
@@ -1423,7 +1423,7 @@ through `-c model=...`, overriding the CLI's configured model for that call.
|
||||
Native review also sets `-c review_model=...` to that selection, overriding any
|
||||
separate review-model pin.
|
||||
|
||||
On Codex hosts, the Claude outside-voice skill is `gstack-claude-code`. Its
|
||||
On Codex hosts, the Claude outside-voice skill is `$claude-code`. Its
|
||||
review, challenge, and consult calls preserve Claude's configured model.
|
||||
`GSTACK_CLAUDE_MODEL=<model>` supplies an explicit override, including resumed
|
||||
sessions; a model named in your request takes precedence. Harness routing is
|
||||
@@ -1470,7 +1470,7 @@ Claude: Running independent Codex review...
|
||||
|
||||
## `/claude-code`
|
||||
|
||||
Claude Code provides the outside reviewer when gstack runs in Codex. Other non-Claude harnesses also expose this skill for explicit requests; Claude Code itself omits it. External harnesses install it as `/gstack-claude-code`.
|
||||
Claude Code provides the outside reviewer when gstack runs in Codex. Other non-Claude harnesses also expose this skill for explicit requests; Claude Code itself omits it. External harnesses install it as `/gstack-claude-code`; Codex lists it as `claude-code`, so mention it as `$claude-code`.
|
||||
|
||||
**Review** supplies the branch diff for a read-only pass/fail review. **Challenge** asks Claude Code to find concrete failure cases in the same diff. **Consult** supports read-only repository exploration and resumes the session saved in `.context/claude-session-id`. Review and challenge receive context from the parent workflow and run without tools; consultation can read and search files.
|
||||
|
||||
|
||||
Reference in new issue
Block a user