@akasecurity/ai-tc-codex 0.9.6 → 0.9.8
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/.codex-plugin/plugin.json +1 -1
- package/package.json +4 -3
- package/scripts/apply-suppressions.js +1653 -1044
- package/scripts/backfill.js +3303 -1385
- package/scripts/dashboard.js +131 -10
- package/scripts/filescan.js +3348 -1421
- package/scripts/firstrun.js +3067 -1143
- package/scripts/intro.js +1162 -932
- package/scripts/onboard.js +1619 -1028
- package/scripts/post-tool-use.js +3373 -1348
- package/scripts/pre-tool-use.js +3191 -1179
- package/scripts/query.js +3071 -1142
- package/scripts/reconcile.js +3230 -1315
- package/scripts/remediate.js +3250 -1332
- package/scripts/scan-worker.js +1114 -916
- package/scripts/session-start.js +3425 -1341
- package/scripts/start-light.js +1080 -851
- package/scripts/stop.js +1380 -902
- package/scripts/sync.js +19413 -0
- package/scripts/user-prompt-submit.js +3362 -1350
- package/skills/setup/SKILL.md +0 -17
package/skills/setup/SKILL.md
CHANGED
|
@@ -828,20 +828,3 @@ pointer inside a Bash command is denied rather than executed or substituted.
|
|
|
828
828
|
If the user asks about recovering a redacted value, point them at the vault
|
|
829
829
|
surfaces that do exist on this machine (`aka vault show`, the dashboard's
|
|
830
830
|
Vault page) rather than implying this plugin can reveal anything.
|
|
831
|
-
|
|
832
|
-
**The model-judge step cannot run on native Windows.** The judge spawns the
|
|
833
|
-
`codex` CLI through a shell-free `execFileSync`, and an npm-installed CLI is a
|
|
834
|
-
`.cmd` shim on Windows, which Node has refused to spawn without a shell since
|
|
835
|
-
the CVE-2024-27980 fix. So a user who grants model-judge consent gets a spawn
|
|
836
|
-
failure at that step rather than a rated set of findings, and the step exits
|
|
837
|
-
non-zero instead of completing. Everything before it is unaffected — the
|
|
838
|
-
calibration questions, the consent-gated historical read, and the local ruleset
|
|
839
|
-
scan all run, and every finding is still detected and redacted locally. What is
|
|
840
|
-
unavailable is only the model pass that rates false positives and severity.
|
|
841
|
-
|
|
842
|
-
Say this plainly if a Windows user asks why setup stopped there, and do not
|
|
843
|
-
offer to retry it: the fix is a code change, tracked separately. Two things
|
|
844
|
-
worth getting right if it comes up — running under WSL2 is Linux rather than
|
|
845
|
-
Windows and is not affected, and declining model-judge consent is not a
|
|
846
|
-
workaround so much as the other supported path, since the local scan is what
|
|
847
|
-
produces the findings in the first place.
|