codex-usage-tracker 0.18.0

What's Changed

• docs: warn about compacted tool output by @douglasmonsky in #135
• Add maintained hardening gates by @douglasmonsky in #136
• Normalize Python formatting by @douglasmonsky in #137
• Define explicit Tach modules by @douglasmonsky in #138
• Fix GPT-5.6 pricing and credit compatibility by @douglasmonsky in #139
• Prepare 0.17.2 release by @douglasmonsky in #140
• ci: maintain a clean Pyright surface by @douglasmonsky in #141
• chore: refresh file length ratchet baseline by @douglasmonsky in #142
• refactor: enforce package domain boundaries by @douglasmonsky in #143
• chore: enforce dependency and dead code gates by @douglasmonsky in #144
• chore: enforce workflow and dependency security by @douglasmonsky in #145
• chore: ratchet reviewed Bandit findings by @douglasmonsky in #151
• chore: enforce whole-source Pyright by @douglasmonsky in #152
• refactor: isolate MCP dogfood jobs by @douglasmonsky in #153
• refactor: split MCP allowance tools by @douglasmonsky in #154
• refactor: split MCP discovery tools by @douglasmonsky in #155
• refactor: split MCP investigation tools by @douglasmonsky in #156
• refactor: split MCP dashboard tools by @douglasmonsky in #157
• refactor: split report discovery family by @douglasmonsky in #158
• refactor: isolate hypothesis classification by @douglasmonsky in #159
• refactor: split report query family by @douglasmonsky in #160
• refactor: split hypothesis evaluators by @douglasmonsky in #161
• refactor: split agentic report family by @douglasmonsky in #162
• refactor: split action brief reports by @douglasmonsky in #163
• refactor: split investigation walk reports by @douglasmonsky in #164
• refactor: split content index queries by @douglasmonsky in #165
• refactor: split content index parsing by @douglasmonsky in #166
• refactor: split content index persistence by @douglasmonsky in #167
• refactor: split allowance statistics by @douglasmonsky in #168
• refactor: split server dashboard pages by @douglasmonsky in #169
• refactor: split diagnostic families by @douglasmonsky in #170
• refactor: split CLI command families by @douglasmonsky in #171
• chore: complete hardening roadmap by @douglasmonsky in #172
• docs: add complexity hardening roadmap by @douglasmonsky in #173
• refactor: simplify agentic evidence by @douglasmonsky in #174
• refactor: simplify agentic report assembly by @douglasmonsky in #175
• refactor: simplify content indexing flow by @douglasmonsky in #176
• chore: upgrade Agent Maintainer frontend governance by @douglasmonsky in #178
• feat: complete final dashboard redesign by @douglasmonsky in #179
• fix: revalidate React dashboard assets by @douglasmonsky in #217
• fix: use scope totals in overview metrics by @douglasmonsky in #218
• fix: harden full-scope dashboard loading by @douglasmonsky in #219
• chore: prepare compression lab delivery by @douglasmonsky in #220
• feat: add compression attribution foundation by @douglasmonsky in #221
• chore(deps): bump actions/setup-python from 5 to 6 by @dependabot[bot] in #147
• chore(deps): bump actions/setup-node from 4 to 6 by @dependabot[bot] in #148
• chore(deps): bump actions/checkout from 4 to 7 by @dependabot[bot] in #149
• chore(deps): bump actions/upload-artifact from 4 to 7 by @dependabot[bot] in #150
• chore(deps): bump actions/setup-go from 5 to 6 by @dependabot[bot] in #146
• feat: build compression analysis profiles by @douglasmonsky in #222
• refactor: reduce compression cold-build overhead by @douglasmonsky in #224
• feat: stream compression evidence rows by @douglasmonsky in #226
• refactor: compact compression detector evidence by @douglasmonsky in #227
• test: benchmark normalized compression evidence by @douglasmonsky in #228
• feat: persist detector-ready compression facts by @douglasmonsky in #231
• feat: add compression revision state by @douglasmonsky in #233
• perf: optimize compression candidate publication by @douglasmonsky in #235
• perf: parallelize first-refresh ingestion by @douglasmonsky in #237
• Improve Simplified Chinese dashboard localization by @LH-03 in #229
• feat: add async compression MCP surface by @douglasmonsky in #239
• feat: add overlap-aware compression simulator by @douglasmonsky in #241
• fix: preserve dashboard module progress by @douglasmonsky in #245
• perf: baseline dashboard query routes by @douglasmonsky in #246
• refactor: add derived fact refresh hook by @douglasmonsky in #247
• feat: materialize recommendation facts by @douglasmonsky in #248
• refactor: wire recommendation-aware refresh by @douglasmonsky in #249
• refactor: route recommendations through indexed facts by @douglasmonsky in #250
• perf: materialize recommendation thread summaries by @douglasmonsky in #251
• refactor: centralize dashboard query identity by @douglasmonsky in #252
• refactor: share dashboard module progress state by @douglasmonsky in #253
• refactor: preserve dashboard source identity by @douglasmonsky in #254
• refactor: migrate cache and reports queries by @douglasmonsky in #255
• refactor: migrate investigator queries by @douglasmonsky in #256
• refactor: migrate diagnostics query lifecycle by @douglasmonsky in #257
• feat: run diagnostic refreshes as dashboard jobs by @douglasmonsky in #258
• feat: add persistent Compression Lab dashboard jobs by @douglasmonsky in #261
• feat: harden dashboard query caching by @douglasmonsky in #262
• refactor: enforce dashboard query pipeline contracts by @douglasmonsky in #263
• docs: complete dashboard query pipeline roadmap by @douglasmonsky in https://github.com/...

codex-usage-tracker 0.17.1

0.17.1

Patch release for the post-0.17.0 dogfood pass.

• Stabilizes async dogfood cache fingerprints so repeated unchanged MCP dogfood runs can reuse cached reports even when SQLite file metadata changes during read/report activity.
• Keeps dogfood cache invalidation tied to indexed usage/content signatures and relevant config files.

Validation: PR checks passed, TestPyPI and PyPI publish workflows passed, fresh local pipx upgrade reports codex-usage-tracker 0.17.1.

codex-usage-tracker 0.17.0

0.17.0

• Add agentic MCP investigation tools for hypothesis-driven usage diagnostics, compact evidence briefs, and actionable recommendation reports.
• Add repeated file rediscovery, shell churn, and large low-output usage diagnostics for avoidable workflow waste.
• Improve bundled skill guidance for routing usage questions into tracker endpoints, Headroom suggestions, and custom remediation ideas.
• Add async dogfood progress polling and warm result caching for longer investigations.
• Speed up content indexing with incremental append metadata, parallel source parsing, and batched FTS rebuilds.
• Add dashboard refresh progress bars, no-cap paged loading, loaded-call metrics, token breakdowns, persisted loading preferences, and restored usage-drain chart data.

Validation: release workflow passed and published codex-usage-tracking==0.17.0 to TestPyPI and PyPI.

codex-usage-tracker 0.15.1

0.15.1

This is the stable public release for the 0.15 usage-coaching work.

Highlights

• Companion plugin/skills now treat token-waste discovery as diagnosis plus remediation, including Headroom suggestions when available, dashboard verification paths, and custom local automation ideas.
• README now centers the conversational plugin/skill workflow and links synthetic aggregate-only example conversation docs.
• React dashboard package resources are fixed in the PyPI wheel, including the tracked React index.html entrypoint required by serve-dashboard.
• Calls/Details screenshots and dashboard smoke tests were refreshed for the showcase branch.

Validation

• GitHub CI passed across Python 3.10-3.14.
• PyPI Trusted Publishing workflow completed successfully.
• Public install smoke passed for codex-usage-tracking==0.15.1.

Note: 0.15.0 was uploaded first but its wheel missed the React dashboard index.html resource. Use 0.15.1.

codex-usage-tracker 0.13.1

Patch release adding guided usage-summary diagnostics across CLI, API, and dashboard, with release metadata aligned to 0.13.1.

codex-usage-tracker 0.12.1

Patch release for the package-domain boundary split and release hardening.

Highlights:

• Ships the package-domain boundary refactor behind compatibility facades.
• Splits allowance pricing helpers while preserving public imports.
• Fixes the installed console-script target after the CLI package split.
• Updates package, plugin, and MCP runtime pins to 0.12.1.

Validation:

• Local release gate passed, including ruff, mypy, pytest, coverage, compileall, dashboard JS syntax, release checks, build/twine, dist inspection, and installed-package smoke.
• PR #79 passed CI before squash merge.
• TestPyPI and PyPI installs both reported codex-usage-tracker 0.12.1.

codex-usage-tracker 0.12.0

Summary

0.12.0 is the maintainability repair release. It keeps the public CLI/API behavior stable while reducing the internal module size and coupling behind the dashboard, diagnostics, parser/context, store/query, and usage-drain report paths.

Highlights

• Refactored CLI, dashboard server routing, SQLite store/query boundaries, context parsing, diagnostics, and usage-drain modeling into smaller modules with compatibility facades.
• Added Agent Maintainer ratchet configuration, strict local tach check, maintainability boundary docs, and a maintainability scorecard.
• Hardened Diagnostics dashboard behavior around projected weekly credits charts, command expansion, live row status, and desktop/mobile smoke coverage.
• Cleaned up warning-sensitive tests so the suite runs without SQLite/resource-handle ResourceWarning noise.
• Validated the refactor against a frozen local JSONL rebuild comparison before release.

Validation

• GitHub CI passed on Python 3.10 through 3.14.
• Local release gate passed: Ruff, mypy, pytest, coverage, compileall, dashboard JS syntax, release checks, build, twine, dist checks, and installed-package smoke.
• TestPyPI publish and clean venv install verified for codex-usage-tracking==0.12.0.

codex-usage-tracker 0.11.4

Summary

Patch release for large local histories.

Fixes

• Fix codex-usage-tracker setup on large Codex histories by batching stale diagnostic-fact deletes so SQLite does not exceed bound-variable limits during usage index refresh.
• Align package, plugin, and MCP launcher metadata to 0.11.4, including the MCP runtime cache version.

Tests

• Local release gate passed: ruff, mypy, pytest, coverage, compileall, dashboard JS syntax, release checker, build, twine, and dist checker.
• PR #71 CI passed across Python 3.10, 3.11, 3.12, 3.13, 3.14, and package build.

Refs #69

codex-usage-tracker 0.11.3

Summary

Patch release for Windows live dashboard startup.

Fixes

• Force safe MIME types for served dashboard .js, .css, .json, and .svg assets. This avoids Windows registry mappings that can make Python serve .js as text/plain, which Chrome blocks when nosniff is enabled.
• Adds regression coverage that simulates the Windows text/plain MIME path.

Validation

• GitHub Actions CI passed for PR #68.
• TestPyPI publish and clean install smoke passed.
• PyPI publish and clean install smoke passed.

Thanks @Vadevious for the detailed Windows repro and diagnosis.

PR: #68
PyPI: https://pypi.org/project/codex-usage-tracking/0.11.3/

codex-usage-tracker 0.11.2

Patch release for served-dashboard hydration.\n\n- Fixes #66: live dashboard shells with zero embedded rows now hydrate from /api/usage when indexed rows are available.\n- Adds regression coverage for serve-dashboard --no-refresh empty-shell hydration.\n- Includes 0.11.2 CLI/plugin version bumps and release gate hardening.\n\nPR: #67\n\nThanks @Vadevious for the detailed Windows repro and root-cause analysis in #66.