You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: README.md
+7-7Lines changed: 7 additions & 7 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -14,15 +14,15 @@ codex --full-auto
14
14
15
15
`codex --full-auto` is the recommended default once this wizard is installed: you keep the repo guardrails and hook enforcement, but day-to-day editing and runs stay low-friction. Use plain `codex` instead if you want more manual confirmation.
16
16
17
-
Bare `npx codex-sdlc-wizard` is the adaptive interactive path. It bootstraps the repo-local guardrails first, then hands off into a live Codex setup session so the unresolved setup questions happen inside Codex instead of inside a shell checklist. `setup --yes` still exists for automation, but it is not the normal human path.
17
+
Bare `npx codex-sdlc-wizard` is the adaptive interactive path. It bootstraps the repo-local guardrails first, then hands off into a live plain Codex setup session so the unresolved setup questions happen inside Codex instead of inside a shell checklist. At that first-run handoff prompt, press Enter for plain `codex` or type `full-auto` if you explicitly want `codex --full-auto`. `setup --yes` still exists for automation, but it is not the normal human path.
Copy file name to clipboardExpand all lines: ROADMAP.md
+13-12Lines changed: 13 additions & 12 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -2,7 +2,7 @@
2
2
3
3
## Current State
4
4
5
-
-`codex-sdlc-wizard@0.7.9` and `v0.7.9` are the current release target for the setup/update scope-control stabilization patch
5
+
-`codex-sdlc-wizard@0.7.10` and `v0.7.10` are the current release target for the first-run plain-Codex handoff patch
6
6
- npm trusted publishing is configured and the GitHub release workflow is now proven for real OIDC publish
7
7
- the repo now ships both a Codex skill package (`SKILL.md`, `agents/openai.yaml`) and the installer/setup adapter (`install.sh`, `setup.sh`)
8
8
- the npm CLI now defaults to adaptive interactive setup instead of requiring an explicit `setup` subcommand for the main human path
@@ -19,51 +19,52 @@
19
19
- setup/update guidance now treats verification as diagnostic for product failures and stops before editing application code or application tests without explicit user consent
20
20
- setup/update guidance now tells users to exit and reopen Codex after hook/skill repairs, without rerunning setup/update just for that restart
21
21
- install/setup/update now write and repair repo-local `.codex/config.toml` model keys for the selected profile, while preserving unrelated MCP, sandbox, approval, and custom config
22
+
- first-run live setup now defaults to plain `codex` after bootstrap and requires an explicit `full-auto` choice to start that setup handoff with `codex --full-auto`
22
23
- the repo now ships a consumer bug-report template for install/setup/runtime failures
23
24
- the public README now leads with the real `@latest` adaptive setup path and keeps the top section consumer-focused
24
25
- benchmark and pilot-rollout ledgers now exist so model/default-use decisions can be measured, not guessed
25
26
- release, packaging, npm, skill, setup, adapter, update, and E2E tests are green when the parity merge is complete
26
27
27
28
## Next Release Cycle
28
29
29
-
### 0.7.10
30
+
### 0.7.11
30
31
31
-
Purpose: prove the post-`0.7.9` consumer path on real repos and stabilize any reusable wizard bugs without changing the default-use claim early.
32
+
Purpose: prove the post-`0.7.10` consumer path on real repos and stabilize any reusable wizard bugs without changing the default-use claim early.
32
33
33
34
Scope:
34
-
- run `0.7.9` on 3-5 pilot repos and log results in `benchmarks/pilot-rollout.csv`
35
+
- run `0.7.10` on 3-5 pilot repos and log results in `benchmarks/pilot-rollout.csv`
35
36
- cut a stabilization patch only if pilots surface another reusable wizard bug
36
37
- keep the default-use recommendation gated on the measurable pilot summary
37
-
- keep the separate model experiment running, but do not let it block pilot rollout work
38
+
- keep separate model-profile measurement running, but do not let it block pilot rollout work
38
39
39
40
## Tracker Cleanup
40
41
41
42
The issue tracker is currently clear.
42
43
43
44
- open a new issue only when pilot consumption exposes a proven reusable wizard bug
44
-
- avoid speculative backlog churn while `0.7.9` is being consumed on real repos
45
+
- avoid speculative backlog churn while `0.7.10` is being consumed on real repos
45
46
46
47
## Remaining Backlog
47
48
48
-
After `0.7.9`, the main backlog is:
49
+
After `0.7.10`, the main backlog is:
49
50
50
51
- pilot rollout proof for default use on real repos
51
52
- any reusable wizard fixes discovered during the pilot set
52
-
- model experiment data collection for `mixed` vs `maximum`
53
+
- model-profile measurement data collection for `mixed` vs `maximum`
53
54
- top-level proof-run parallelization to reduce release-wall-clock time without weakening suite coverage
54
55
- later creator-tool research after the active backlog stays under control
55
56
56
57
## Working Order
57
58
58
-
1. Prove the default-use gate on 3-5 pilot repos with `0.7.9`
59
-
2. Ship `0.7.10` only if pilot rollout surfaces another reusable wizard bug
59
+
1. Prove the default-use gate on 3-5 pilot repos with `0.7.10`
60
+
2. Ship `0.7.11` only if pilot rollout surfaces another reusable wizard bug
60
61
3. Keep creator-tool investigation behind the active backlog
61
62
62
63
## Default-Use Gate
63
64
64
65
Before calling this the default Codex SDLC path, prove it on real pilot repos instead of just repo-self-tests.
65
66
66
-
- run `0.7.9` on 3-5 pilot repos
67
+
- run `0.7.10` on 3-5 pilot repos
67
68
- require pilot success >= 95% before default use
68
69
- allow no more than 1 reusable wizard bug across the pilot set
69
70
- track the pilot set in `benchmarks/pilot-rollout.csv`
@@ -75,7 +76,7 @@ After the current backlog is under control, investigate whether Codex's built-in
75
76
76
77
- evaluate `Skill Creator` as a possible future aid for skill-structure maintenance
77
78
- evaluate `Plugin Creator` only as later research, since plugins are not part of the current shipping path
78
-
-experiment with`gpt-5.4-mini` for the main working pass while keeping `xhigh` for review or cross-model review, and compare that against simply running the whole slice at `xhigh`
79
+
-measure`gpt-5.4-mini` for the main working pass while keeping `xhigh` for review or cross-model review, and compare that against simply running the whole slice at `xhigh`
79
80
- if the mixed mode proves out, add an easy toggle between two explicit profiles:
80
81
-`mixed`: `gpt-5.4-mini` for the main pass plus `xhigh` review
81
82
-`maximum`: `gpt-5.4` / `xhigh` for the whole slice as the "ultimate mode"
Copy file name to clipboardExpand all lines: SKILL.md
+1-1Lines changed: 1 addition & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -17,7 +17,7 @@ Use the bundled scripts like this:
17
17
1. If the user wants the simplest install, run the bundled `install.sh` from this skill bundle against the current working repo.
18
18
2. If the user wants repo-aware setup or regenerated docs, run the bundled `setup.sh --yes` from this skill bundle against the current working repo.
19
19
3. Tell the user exactly which path you chose: `install.sh` for baseline enforcement, `setup.sh` for adaptive setup.
20
-
4. After installation, tell the user to start a fresh Codex session so hooks and repo docs are loaded cleanly. Recommend `codex --full-auto` as the default start mode once the guardrails are installed, and mention plain `codex` as the manual fallback.
20
+
4. After installation, tell the user to start a fresh Codex session so hooks and repo docs are loaded cleanly. Recommend plain `codex` for the first live setup handoff unless the user explicitly chooses `full-auto`; recommend `codex --full-auto` as the default start mode after setup is complete and the guardrails are loaded. Mention plain `codex` as the manual fallback for daily work.
21
21
The fresh session should also pick up the repo-scoped `\$sdlc` skill under `.agents/skills`. Repo-scoped skill coverage is still a work in progress; `gdlc` and `rdlc` are planned next.
22
22
For setup/update bootstrap work, recommend the `maximum` profile via `--model-profile maximum` as the safer default. For routine work after bootstrap, point users back to the `mixed` profile via `--model-profile mixed` for the better speed / lower latency / lower token path with `xhigh` review.
23
23
This repo stays on `maximum`; when maintaining `codex-sdlc-wizard` itself, keep the wizard repo on the stability-first path because the work is unusually meta.
0 commit comments