Skip to content

Explain T3 usage scopes and account limits - #37

Merged
gustavonline merged 1 commit into
mainfrom
qual/factory31-skill-read-20261013
Oct 15, 2026
Merged

gustavonline merged 1 commit into
mainfrom
qual/factory31-skill-read-20261013

Conversation

@gustavonline

Copy link
Copy Markdown
Contributor

Foundation now explains how to read T3's Usage view without treating an API estimate as a subscription bill or an environment filter as project accounting.

Previously the guide only pointed users to T3. It now explains, for example, why another environment using the same provider history does not create a separate project ledger or account allowance. The existing adoption record gains the relevant source, scope and coverage fields.

Checks: eight skills/zero errors, 11 passing unit tests and nine staged Foundation files with matching hashes. An existing Mac operator checked the actual nightly3010 views and filters; independent Sol Max review accepted the exact candidate. Project attribution, complete account-history coverage and independently fresh quota data remain unavailable or unverified.

Refs #35. Existing upstream request.

Review and evidence scope
  • Five changed files; Foundation version0.1.14. Independent native Codex gpt-6.1-sol Max/Standard R2: ACCEPT SOURCE, no blocking source findings.
  • Exact reviewed patch SHA256: c82b1305fb0d4de1d1070089c39d2b2ddc13570dbaec8f70cf67a02211b40766.
  • The last candidate delta is only the qualification row incorporating the operator's real UI observation. The checker/whitespace checks are fresh; unit/staging evidence is carried on identical relevant bytes, not counted as new runs.
  • The first review attempt stopped before source inspection; it produced no verdict. Its obsolete permission request is never an execution authorization. The delivery record supersedes the stale review wording in the earlier check receipt without altering that receipt.
  • The operator observed Cost/Tokens/Limits, history windows, provider/model views and the explicit API-estimate/unpriced labels. No project/thread/account filter was observed. Limits had no clear as-of timestamp; no forced refresh or reset was performed.
  • No environments, accounts, pricing overrides, permissions, runtime or installed skills changed. Staging and source acceptance are separate from installation and native isolation.
  • Previous releasev0.1.13 remains the recovery source. Private raw native/check evidence is retained with the installation. Only this bounded guidance is being delivered; the custom#32/Qualify safe native T3 maintenance and stable-channel adoption #34 candidates stay parked.

Expand Foundation's usage note into a practical native procedure. Separate
recorded tokens, API estimates, subscription limits and actual charges, and
explain why another environment sharing provider history is not project
accounting or an OS boundary. Record the attributed native UI observation
and the remaining upstream attribution/freshness limits.

Refs #35

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@gustavonline
gustavonline merged commit 9f20220 into main Oct 15, 2026
1 check passed
@gustavonline
gustavonline deleted the qual/factory31-skill-read-20261013 branch October 15, 2026 15:10
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant