From 975d29ce2c6f8e4d75cdc9e05b11cefd36d0626f Mon Sep 17 00:00:00 2001 From: Peter Bons Date: Wed, 28 Oct 2026 14:57:53 +0100 Subject: [PATCH] Add an anonymized example HTML report and link it from the README Co-Authored-By: Claude Opus 5.5 --- README.md | 4 + example-output.html | 1686 +++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 1690 insertions(+) create mode 100644 example-output.html diff --git a/README.md b/README.md index 342ebc5..1c7974e 100644 --- a/README.md +++ b/README.md @@ -22,6 +22,10 @@ npx ccreports@latest --source copilot # GitHub Copilot instead ( Or install it globally: `npm install -g ccreports`, then run `ccreports`. +See [example-output.html](example-output.html) for an example HTML report: per-session usage +across Claude Code, GitHub Copilot and Cursor, with adoption and recommendations +(session titles and project paths anonymized). + ## Install as a Claude Code plugin At the prompt of a Claude Code terminal session: diff --git a/example-output.html b/example-output.html new file mode 100644 index 0000000..5b1ee7b --- /dev/null +++ b/example-output.html @@ -0,0 +1,1686 @@ + + + + + +Claude Code, GitHub Copilot and Cursor per-session usage + + + +
+

Claude Code, GitHub Copilot and Cursor per-session usage

+

Generated 2026-10-28 14:50 with ccreports 0.5.2 · GitHub Copilot cost from its billed AI credits · other cost estimates at Claude API list prices

+
  • Most of your sessions are Chat (35% of 197).
  • 49% of sessions work agent-first or with multiple agents.
  • Agent-first or multi-agent: 52% of your earlier sessions, 45% of your recent ones.
  • PRs per session: Assisted 0.56, Agent-first 0.49, Multi-agent 0.35.
  • Cursor: No clear improvement found in this period - your usage already follows the patterns checked.
+
+
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
SessionProjectLast ActivityModelsInputOutputCache CreateCache ReadTotal TokensCost (USD)Max ContextAvg ContextToolsSkillsCmdsAgents
fc25ea3cZ:\Sources\Fabrikam.Portal2026-02-07 12:1900000$0.000034031
116117f8Z:\Sources\Fabrikam.Portal2026-02-07 12:4000000$0.000037110
acc915cdZ:\Sources\Fabrikam.Portal2026-02-07 13:5700000$0.000064220
ecb9f565Z:\Sources\Fabrikam.Portal2026-02-07 14:1600000$0.000029110
f0045adeZ:\Sources\Fabrikam.Portal2026-02-07 14:2100000$0.000017110
37c2d3b8Z:\Sources\Fabrikam.Portal2026-02-07 14:5200000$0.000031110
cf0685c5Z:\Sources\Fabrikam.Portal2026-02-07 16:5900000$0.000043120
2f96dc3cZ:\Sources\Fabrikam.Portal2026-02-08 09:3800000$0.000038101
e4ca8144…brikam.Estimates.Web\Fabrikam.Estimates2026-02-08 16:3100000$0.000045001
5e38e2f6C:\Users\jdoe2026-02-19 14:4100000$0.00001000
61b7790dZ:\2026-02-20 08:5400000$0.000012330
fb37e7cdZ:\Sources\Fabrikam.Estimates.Core2026-02-20 16:1400000$0.000015110
4c33841fZ:\Sources\Fabrikam.Estimates.Core2026-02-21 15:4200000$0.0000157411
e2642863Z:\Sources\Fabrikam.Estimates.Core2026-02-22 13:3100000$0.0000174100
033aaafdZ:\Sources\TripMate2026-03-14 19:0500000$0.00000010
8b76efb3Z:\Sources\TripMate2026-03-16 08:09sonnet-4.6, gpt-5.3-codex, haiku-4.53,705,367295,519054,183,68058,184,566$28.5200906191910
cec3f394Z:\Sources\Sandbox2026-03-17 09:5700000$0.00003200
2a0b3994Z:\Sources\Sandbox2026-03-17 10:0100000$0.00002000
e87ef260Z:\Sources\Sandbox2026-03-17 10:3300000$0.00001010
32fa8386Z:\Sources\Sandbox2026-03-17 10:3500000$0.000010010
121b6357Z:\Sources\Sandbox2026-03-17 13:5800000$0.00000000
19d908f9Z:\Sources\Fabrikam.Portal2026-03-17 14:02sonnet-4.624,9809050117,724143,609$0.12006000
564096a1Z:\Sources\TripMate2026-03-18 15:4500000$0.0000238653
b067b3d8Z:\Sources\TripMate2026-03-18 21:30gpt-5.4, opus-4.6, sonnet-4.5, haiku-4.5747,00572,116018,737,59719,556,718$6.7200323542
53351165Z:\Sources\Fabrikam.Portal2026-03-21 19:25haiku-4.5, gpt-5.4141,86218,8350959,8641,120,561$0.260099002
82064b8fZ:\Sources\Fabrikam.Portal2026-03-22 19:58gpt-5.4, haiku-4.5136,05921,49301,303,8851,461,437$0.080095111
70f5e245Z:\Sources\Fabrikam.Estimates.Core2026-03-23 09:0200000$0.00004100
81a0274fZ:\Sources\Fabrikam.Estimates.Core2026-03-23 13:09gpt-5.4, sonnet-4.5, haiku-4.5320,76953,57804,217,5654,591,912$0.6000215733
ffa7cb78Z:\Sources\Ledger2026-03-26 11:48gpt-5.4, haiku-4.5, sonnet-4.51,027,438121,263022,475,25023,623,951$1.3300595533
b68915b5Z:\Sources\Ledger2026-03-26 14:45gpt-5.4, haiku-4.5159,88142,68701,520,8801,723,448$0.1300109101
8c0b38c2Z:\Sources\TripMate2026-04-08 08:52gpt-5.4935,193130,766023,223,29624,289,255$0.0000429441
2f8fde59Z:\Sources\Fabrikam.Estimates.Web2026-04-08 11:04gpt-5.4404,95812,03802,347,5202,764,516$0.000067000
4466b04dZ:\Sources\BudgetBuddy2026-04-11 20:56gpt-5.4816,11569,36808,624,3849,509,867$0.0000209550
b148b9feZ:\Sources\BudgetBuddy2026-04-12 20:46gpt-5.426,5502,6430105,088134,281$0.000011000
e8621356Z:\Sources\TripMate2026-04-13 16:1800000$0.000042000
3356b236Z:\Sources\TripMate2026-04-14 10:2800000$0.0000212430
bf832b37Z:\Sources\TripMate2026-04-15 10:0600000$0.0000422560
dd47d79bZ:\Sources\TripMate2026-04-15 16:0100000$0.0000543860
2c665b05Z:\Sources\TripMate2026-04-15 21:4400000$0.0000298660
1ef6d2c1Z:\Sources\TripMate2026-04-16 08:51gpt-5.444,3662,4870192,128238,981$0.000020000
69fbcc76Z:\Sources\TripMate2026-04-17 09:4500000$0.000062220
1f498a8cZ:\Sources\TripMate2026-04-17 09:5800000$0.000089110
Refactor order service retriesZ:\Sources\TripMate2026-04-30 20:1100000$0.0000296440
Add dark mode toggleZ:\Sources\TripMate2026-05-02 00:03gpt-5.41,447,715186,011027,268,48028,902,206$0.0000440660
Invoice PDF exportZ:\Sources\TripMate2026-05-10 19:53gpt-5.4317,36037,69103,645,6964,000,747$0.0000124220
Fix flaky login testZ:\Sources\TripMate2026-05-12 10:17gpt-5.447,4612,7270241,792291,980$0.000022000
Migrate CI to new runnersZ:\Sources\TripMate2026-05-12 10:22gpt-5.426,7877,4460105,472139,705$0.000025010
Pagination for search resultsZ:\Sources\TripMate2026-05-17 12:11gpt-5.4226,13248,49208,460,1608,734,784$0.0000284550
Webhook signature validation…\Users\jdoe\AppData\Temp\178023722369232026-05-31 16:39gpt-5.425,9131,375077,440104,728$0.00006000
Cache warmup on startup…\Users\jdoe\AppData\Temp\178023722369232026-05-31 19:41gpt-5.430,0003,6060247,040280,646$0.000013000
Rate limiter for public APIZ:\Sources\TripMate2026-06-01 09:10gpt-5.4404,93464,22004,749,8245,218,978$0.0000275220
Upgrade to .NET 10Z:\Sources\TripMate2026-06-02 14:36gpt-5.4109,3397,6050482,816599,760$0.510045000
Dockerfile cleanupZ:\Sources\Workshop2026-06-21 11:04gpt-5.459,5798,5040513,024581,107$0.4000108000
Add OpenTelemetry tracingZ:\Sources\TripMate2026-06-28 12:10gpt-5.4417,79774,655011,174,40011,666,852$4.9600148220
Weekly digest email jobC:\Users\jdoe2026-09-05 14:18gpt-5.418,3151200018,435$0.05000000
Inventory sync bugC:\Users\jdoe2026-09-05 15:38qwen3.5:9b39,6561380039,794$0.00000000
Inventory sync bugZ:\Sources\Samples.Scripts\ai2026-09-16 17:05gpt-5.4163,8492,1530999,9361,165,938$0.690029100
43380c17C:\Users\jdoe2026-09-17 14:28sonnet-5121,61357,052460,731519,408$0.3495,72086,2994010
CSV import validation…b71f-29c3ce0625d9\scratchpad\copilotcwd2026-09-27 17:02gpt-5.428,004779061,95290,735$0.10008001
Add health check endpointZ:\Sources\Samples.Scripts\ai2026-09-27 20:24gpt-5.443,7075,944089,088138,739$0.22008000
Feature flag for checkout v2Z:\Sources\Samples.Scripts\ai2026-09-27 21:17gpt-5.453,56114,1390415,232482,932$0.450030000
Review pull request 212Z:\Sources\Samples.Scripts\ai2026-09-27 22:15gpt-5.413,74295011,77625,613$0.04001000
Split monolith into modulesZ:\Sources\Samples.Scripts\ai2026-09-28 08:56gpt-5.413,108660013,174$0.03000000
Postgres index tuningZ:\Sources\Nimbus2026-09-28 13:30opus-5-519056,866210,80813,289,44413,557,308$5.48200,651143,62295110
Explain the auth flow…d9-73f0cbb6e868\scratchpad\test\copilot2026-09-28 14:57gpt-5.414,578134023,55238,264$0.04002000
Write unit tests for parserZ:\Sources\Samples.Scripts2026-09-29 15:08opus-5-58614,738100,1772,924,1833,039,184$1.6897,06072,01133000
Localization for German UIZ:\Sources\Samples.Scripts\ai2026-09-29 16:02gpt-5.447,1496,788060,928114,865$0.230012000
d63727c7C:\Users\jdoe2026-09-30 09:24sonnet-5-5244,99084,502709,740799,256$0.4684,97866,18912000
Fix timezone handling in reportsZ:\Sources\Fabrikam.Portal2026-10-01 16:31opus-5-5912148,5342,436,32160,319,33762,905,104$28.89339,277210,6545410010
Commit, PR, and mergeZ:\Sources\harbor2026-10-02 10:02opus-5-5204,21026,145475,368505,743$0.3956,46150,1538000
Commit, PR, and mergeZ:\Sources\atlas-workspace2026-10-02 10:04opus-5-5244,16224,105580,273608,564$0.3954,42150,36710010
b58e6ad5Z:\Sources\Nimbus2026-10-03 16:46opus-5-5, haiku-4-573469,659589,6229,154,7279,814,742$5.62170,198106,616208244
Commit and push hotfixZ:\Sources\Nimbus2026-10-03 17:00opus-5-562,65126,049137,616166,322$0.2956,15954,5571010
Release notes for 3.4Z:\Sources\Nimbus2026-10-03 18:24opus-5-56039,841212,0945,478,2155,730,210$3.59251,061189,67951010
Cleanup unused NuGet packagesZ:\Sources\Nimbus2026-10-03 19:03opus-5-5, haiku-4-5, sonnet-5-537665,784416,2786,667,4377,149,875$4.02178,197118,498149114
f48cef5dZ:\Sources\Nimbus2026-10-03 19:57opus-5-5, sonnet-5-526098,088399,00221,270,22321,767,573$9.04257,923173,006153211
bf1e1401Z:\Sources\Nimbus2026-10-03 21:02opus-5-5, sonnet-5-5336130,434791,52241,530,41142,452,703$16.73415,632266,235185431
4c59668aZ:\Sources\Nimbus2026-10-05 13:52opus-5-5429125,06482,943108,302$0.2255,17454,0061120
Kubernetes manifest reviewZ:\Sources\Nimbus2026-10-05 19:08haiku-4-5, opus-5-5, sonnet-5-51,36688,678909,19913,656,55314,655,796$7.90221,179123,742291226
f935fbffZ:\Sources\Nimbus2026-10-05 20:35opus-5-5, sonnet-5-5564242,3231,192,86495,765,84297,201,593$32.68609,690362,428338621
Terraform state migrationZ:\Sources\atlas-workspace2026-10-06 09:42fable-5-1, opus-5-52,552404,8017,831,895377,985,613386,224,861$214.08785,048511,1391,8150213
GraphQL schema designZ:\Sources\harbor2026-10-06 11:10opus-5-569866,009832,06340,643,89741,542,667$14.08186,418124,638362005
Map view for store locatorZ:\Sources\atlas-workspace2026-10-06 11:31opus-5-51,114327,5241,790,061128,515,728130,634,427$43.07652,057349,226766115
Offline mode for mobile appZ:\Sources\atlas-workspace2026-10-08 09:13opus-5-51,000178,2401,789,35551,279,44353,248,038$24.17234,374160,9777440110
Push notification settingsZ:\Sources\harbor2026-10-08 09:15opus-5-51,10882,6721,458,71947,901,46049,443,959$20.31250,037155,041576019
Customer portal redesign planZ:\2026-10-08 09:48opus-5-55226,903195,3882,716,7802,939,123$2.64164,687112,00823000
Retry policy for HTTP clientZ:\Sources\harbor2026-10-08 11:38opus-5-51,934160,1431,791,41198,275,289100,228,777$32.91313,427153,5001,0930111
Background job dashboardZ:\Sources\atlas-workspace2026-10-08 11:45opus-5-51,428160,2721,607,25491,706,53693,475,490$30.29265,912183,646807028
Audit log retentionZ:\Sources\Nimbus2026-10-08 13:38opus-5-5, haiku-5-513037,100328,2796,823,3297,188,838$4.23152,138110,02767120
SSO with Entra IDZ:\Sources\Nimbus2026-10-08 14:06opus-5-5, haiku-5-55624,139118,4852,721,6312,864,311$1.74127,416101,43527120
Fix memory leak in workerZ:\Sources\Nimbus2026-10-08 14:46opus-5-57625,506150,3693,438,9763,614,927$2.40119,77394,45840110
Benchmark JSON serializersZ:\Sources\Nimbus2026-10-09 09:08opus-5-59227,323176,7954,239,3344,443,544$2.81122,87096,00552110
Onboarding checklist pageZ:\Sources\Nimbus2026-10-09 10:52opus-5-5, haiku-5-5, sonnet-5-520673,999499,4405,904,0296,477,674$2.91161,99891,737169215
Background job dashboardZ:\Sources\atlas-workspace2026-10-09 11:42opus-5-5, fable-5-1, haiku-4-52,670136,227789,48933,889,26434,817,650$20.78318,823191,582187023
Image resize pipelineZ:\Sources\harbor2026-10-09 14:00opus-5-5, fable-5-15,500437,2373,727,327116,961,395121,131,459$72.96336,605169,6581,0870315
Search ranking tweaksZ:\Sources\Nimbus2026-10-09 16:46opus-5-5, sonnet-5-5442173,4321,037,19468,600,23169,811,299$24.78510,240335,763251311
Add Swagger examplesZ:\Sources\Nimbus2026-10-09 16:57opus-5-5, sonnet-5-5, haiku-5-59253,206273,7294,291,3504,618,377$3.03168,456121,66275221
698a17a4Z:\Sources\Nimbus2026-10-09 17:31opus-5-5, sonnet-5-522679,137291,85317,955,35318,326,569$7.11260,577174,248124221
Dependency vulnerability scanZ:\Sources\Nimbus2026-10-10 13:20haiku-5-5, opus-5-58611,174314,7431,862,2782,188,281$0.93109,50783,49865123
Blob storage lifecycle rulesZ:\Sources\Nimbus2026-10-10 13:58opus-5-58622,35271,1903,410,6713,504,299$1.70103,83980,97647010
Queue consumer scalingZ:\Sources\Nimbus2026-10-10 14:36opus-5-5, sonnet-5-5382145,709825,86818,890,40319,862,362$10.35222,232131,826230118
Profile page accessibility fixesZ:\Sources\Nimbus2026-10-10 15:04opus-5-5, sonnet-5-517672,604237,08912,670,87612,980,745$5.60224,237157,02196121
0c5d78f0Z:\Sources\Nimbus2026-10-10 15:34opus-5-5, sonnet-5-5244113,963709,74628,830,21929,654,172$13.30379,924261,627136221
Shopping cart persistenceZ:\Sources\Nimbus2026-10-10 16:18opus-5-5286,62164,8881,016,0441,087,581$0.8599,32877,21114220
Payment provider sandbox setupZ:\Sources\Nimbus2026-10-10 18:52opus-5-5, sonnet-5-5, haiku-5-5370160,105890,94542,753,60143,805,021$17.80418,214254,242232222
Bicep template for Key VaultZ:\Sources\Nimbus2026-10-10 19:03opus-5-5, sonnet-5-5348131,266411,55636,387,17836,930,348$12.54340,293229,168241201
Data model for subscriptionsZ:\Sources\Nimbus2026-10-10 20:18opus-5-5, sonnet-5-5, haiku-5-5250121,879378,28823,698,31224,198,729$9.52313,884201,330147331
Plan: notification service…9-73f0cbb6e868\scratchpad\test\copilot22026-10-11 14:17gpt-5.433,6734,212088,576126,461$0.17007001
Plan: notification service…9-73f0cbb6e868\scratchpad\test\copilot22026-10-11 14:19gpt-5.415,1381,869061,44078,447$0.08004001
Plan: reporting module phase 2Z:\Sources\Samples.Scripts\ai2026-10-11 14:36opus-5-5332177,156655,94339,626,17140,459,602$16.72377,211242,665226110
Plan: billing API v2…ces\Samples.Scripts\ai\multi-agent-demo2026-10-12 09:52gpt-5.480,17127,1910229,376336,738$0.670014000
Plan: billing API v2…ces\Samples.Scripts\ai\multi-agent-demo2026-10-12 10:00gpt-5.490,34029,9110849,920970,171$0.890024000
5416ce16Z:\Sources\Workshop2026-10-12 10:0700000$0.00000030
Architecture overviewC:\Users\jdoe2026-10-12 14:21opus-5-53613,69634,581980,2651,028,578$0.7567,77656,38214100
Codebase walkthrough…ces\Samples.Scripts\ai\multi-agent-demo2026-10-12 15:25opus-5-5506261,1531,104,47339,813,05541,179,187$22.02283,119161,731360000
Summarize open issuesZ:\Sources\Nimbus2026-10-12 15:43opus-5-5, haiku-5-5, sonnet-5-516052,831422,2056,005,0556,480,251$3.39205,299158,465123104
Search ranking tweaksZ:\Sources\Nimbus2026-10-13 09:27opus-5-5, sonnet-5-5506302,6541,757,74890,303,82692,364,734$37.52650,197381,695300311
Compare logging librariesZ:\Sources\Nimbus2026-10-13 10:03opus-5-5, haiku-5-59221,512131,8104,296,7754,450,189$2.05129,02596,27645120
Evaluate vector databasesZ:\Sources\Nimbus2026-10-13 11:04opus-5-5, haiku-5-54211,59087,9161,572,5031,672,051$1.0499,80879,07017120
Prompt template experimentsZ:\Sources\harbor2026-10-13 15:27opus-5-55818,800217,2672,228,8062,464,931$2.56115,30484,34926000
Chatbot handoff to human agentZ:\Sources\atlas-workspace2026-10-13 18:07opus-5-537279,919897,15818,581,39719,558,846$11.37249,797160,053252014
40d2ef1dZ:\Sources\Nimbus2026-10-16 16:2900000$0.00000010
Agent tool calling demoZ:\Sources\Nimbus2026-10-16 19:31opus-5-510624,054202,1854,493,8744,720,219$3.00131,77590,31160020
MCP server for ticket systemZ:\Sources\Nimbus2026-10-16 20:02opus-5-5185,12536,974531,556573,673$0.5070,15863,1727010
LLM response streamingZ:\Sources\Nimbus2026-10-17 17:25opus-5-517053,045170,51511,409,84011,633,570$4.71201,937139,524105110
16dbdbe0Z:\Sources\Nimbus2026-10-17 17:35opus-5-5, haiku-5-5, sonnet-5-514255,733364,1614,399,4944,819,530$2.41138,83992,101122123
Embeddings for product searchZ:\Sources\Contoso.Support.Agents2026-10-17 17:43opus-5-568289,3181,285,39147,350,32348,725,714$18.42277,887167,334343106
082c5009Z:\Sources\Nimbus2026-10-17 18:01opus-5-5, sonnet-5-522483,959288,44217,959,26818,331,893$7.15249,488176,249119111
Token budget for summariesZ:\Sources\Nimbus2026-10-17 20:01opus-5-5, sonnet-5-57821,219139,7032,716,8932,877,893$1.80126,11788,37442111
Evaluate prompt cachingZ:\Sources\Nimbus2026-10-17 20:51opus-5-5, haiku-5-5, sonnet-5-5302107,953532,25016,830,20617,470,711$6.78251,462149,240190324
Fix 500 on order endpointZ:\Sources\Nimbus2026-10-18 11:50opus-5-5, sonnet-5-54015,743104,5621,095,5551,215,900$1.15101,71475,62127101
Evaluate prompt caching….claude\worktrees\feature-notifications2026-10-18 12:18opus-5-5, sonnet-5-513638,765147,9036,866,1127,052,916$3.10143,219110,47184221
Evaluate prompt cachingZ:\Sources\Nimbus2026-10-18 12:44opus-5-5, sonnet-5-5272112,290345,46527,403,30827,861,335$10.05307,236214,692157211
a1ab63eaZ:\Sources\Nimbus2026-10-18 20:32opus-5-5, haiku-5-5, sonnet-5-512837,560335,7064,175,2964,548,690$2.22145,73496,24183113
Debug failing nightly buildZ:\Sources\Nimbus2026-10-18 20:40opus-5-5, haiku-5-5, sonnet-5-524274,206668,24610,101,89910,844,593$4.77233,784158,785180205
54dccb96Z:\Sources\Nimbus2026-10-18 21:00opus-5-5, sonnet-5-522888,014297,36318,963,92119,349,526$7.53267,736182,881174121
8a831c02…\.claude\worktrees\feature-llm-provider2026-10-18 21:10opus-5-5, sonnet-5-524896,115302,34421,269,06721,667,774$8.24278,604182,863163111
Investigate slow startup…\.claude\worktrees\feature-llm-provider2026-10-19 11:48opus-5-5101,80736,730297,943336,490$0.3971,82266,9374010
Investigate high CPU in APIC:\Users\jdoe2026-10-19 11:49opus-5-5102,85733,461212,367248,695$0.3750,43649,1682000
Exception triage from logsZ:\Sources\Nimbus2026-10-19 12:39opus-5-535050,463541,49416,300,10116,892,408$7.35160,660126,417202014
Hello worldZ:\Sources\Nimbus2026-10-19 12:54haiku-5-5, opus-5-5, sonnet-5-529245,965491,60511,188,62411,726,486$3.05168,229115,352177224
Create a sample todo app…\.claude\worktrees\feature-llm-provider2026-10-19 13:11opus-5-5, haiku-5-54012,925103,0461,568,7101,684,721$1.12112,84183,59025220
be90abbbZ:\Sources\Nimbus2026-10-19 13:23opus-5-5, sonnet-5-532289,202311,52427,662,07128,063,119$9.37274,629182,974193111
Generate test fixtures…\.claude\worktrees\feature-llm-provider2026-10-19 15:16opus-5-56221,003148,7412,639,8612,809,667$2.14111,63989,95738010
49add56cZ:\Sources\Nimbus2026-10-19 15:16haiku-5-5682327,848150,439179,116$0.0161,38359,4312120
Run smoke testsZ:\Sources\Nimbus2026-10-19 15:28haiku-5-5, opus-5-5, sonnet-5-518465,274461,7595,215,4485,742,665$2.46165,740106,086166115
b599a7d7Z:\Sources\Nimbus2026-10-19 16:34opus-5-5, sonnet-5-519649,789241,37414,226,53214,517,891$5.41217,418163,098111111
Echo command testZ:\Sources\Nimbus2026-10-19 16:46opus-5-5, haiku-5-5, sonnet-5-520470,172605,8917,718,7008,394,967$3.68196,792132,166134225
3658eefcZ:\Sources\Nimbus2026-10-19 19:45opus-5-5, sonnet-5-5, haiku-5-5640232,495826,94373,580,37774,640,455$24.01490,730297,098404432
Shell command sanity check…\.claude\worktrees\feature-llm-provider2026-10-19 20:33opus-5-5, haiku-5-55415,043112,0491,890,2412,017,387$1.2789,20574,16127220
Scratch test sessionZ:\Sources\Nimbus2026-10-19 20:38opus-5-5, haiku-5-5, sonnet-5-515054,607307,8298,019,8188,382,404$4.07210,486133,25991212
Scaffold Blazor componentZ:\Sources\Nimbus2026-10-20 10:38opus-5-5, sonnet-5-5234107,585334,49519,989,25220,431,566$8.32288,888192,218132221
Sidebar collapse behaviorZ:\Sources\Nimbus2026-10-20 15:10opus-5-5, haiku-5-5, sonnet-5-5718233,3201,020,75070,810,50872,065,296$23.06472,363238,691458635
2b57b9b1C:\Users\jdoe2026-10-20 15:12opus-5-521412,26915,88728,172$0.1028,15828,1580000
Fullscreen chart viewZ:\Sources\Nimbus2026-10-20 17:00opus-5-5, haiku-5-5578296,117864,92475,083,63576,245,254$26.60535,272330,117430333
Remember last selected tenantZ:\Sources\Nimbus2026-10-21 11:06opus-5-5162,95034,899481,004518,869$0.4369,17664,4906000
Keyboard shortcuts helpZ:\Sources\Nimbus2026-10-21 11:43opus-5-5429,81047,5531,492,2361,549,641$0.8882,64573,32521010
Theme color tokensZ:\Sources\Nimbus2026-10-21 19:08opus-5-5, haiku-5-5, sonnet-5-5404137,914923,10135,502,40636,563,825$12.23335,715228,461283414
Settings page validationZ:\Sources\Nimbus2026-10-21 19:34opus-5-5, sonnet-5-5, haiku-5-5948469,0151,850,717133,220,289135,540,969$45.30703,086391,005773438
Multi-tenant database strategyZ:\Sources\Nimbus2026-10-21 20:35opus-5-5, sonnet-5-5300169,767459,19639,534,49340,163,756$14.46410,755280,193191111
Tenant isolation reviewZ:\Sources\Nimbus2026-10-21 21:27opus-5-5, sonnet-5-5, haiku-5-5542154,574871,81857,106,96258,133,896$17.41437,014294,967347233
Read replica routingZ:\Sources\Nimbus2026-10-21 22:23opus-5-5266123,19535,09058,948$0.2158,28758,2870000
Seed data for demo envZ:\Sources\Nimbus2026-10-21 23:27opus-5-5, sonnet-5-5, haiku-5-5706256,2391,160,67988,199,60089,617,224$27.70541,645358,297415336
Export dashboard to Excel…claude\worktrees\feature-history-search2026-10-22 09:31opus-5-5, haiku-5-5, sonnet-5-5928445,9462,025,591128,126,578130,599,043$44.39738,561391,6267624210
c08eea38…claude\worktrees\feature-history-search2026-10-22 11:28sonnet-5-5, opus-5-530061,148555,82815,613,54516,230,821$4.77124,62499,212201224
Weekly usage reportZ:\Sources\Nimbus2026-10-22 12:43opus-5-5, haiku-5-5162,235154,001332,657488,909$0.7575,30069,5254120
92f68e85…-18da-4e3d-9f0d-4510a78cc871\scratchpad2026-10-22 15:58opus-5-5490315,41647,45063,773$0.1532,16831,4351110
04e20d94…-18da-4e3d-9f0d-4510a78cc871\scratchpad2026-10-22 15:58opus-5-581,45415,895111,782129,139$0.1834,31631,9213100
Weekly usage reportZ:\Sources\usage-insights2026-10-23 10:14opus-5-5526266,3952,566,87381,930,21684,764,010$35.13537,319330,069395230
Cost breakdown per teamZ:\Sources2026-10-23 15:09opus-5-59023,917115,5934,526,7034,666,303$1.96129,861103,16454110
Secrets rotation scriptZ:\Sources\NightOwl2026-10-24 16:49opus-5-55258,197151,7832,156,8352,366,867$2.81117,41688,79544000
424edb79…979-9343-00806d9d745a\scratchpad\cctest2026-10-26 10:58opus-5-5634515,87646,85963,086$0.1430,80530,4052001
b0a7a912…979-9343-00806d9d745a\scratchpad\cctest2026-10-26 10:58opus-5-5623113,56549,19963,001$0.1230,80930,4272001
52acfa10…979-9343-00806d9d745a\scratchpad\cctest2026-10-26 10:59opus-5-5628717,99048,57166,854$0.1430,83030,4372001
1d57ffcb…979-9343-00806d9d745a\scratchpad\cctest2026-10-26 11:00opus-5-561,62734,22248,54284,397$0.2531,21230,6142001
Git history cleanup…343-00806d9d745a\scratchpad\hk-consumer2026-10-26 11:37sonnet-5222,29839,478360,209402,007$0.190010000
Resolve merge conflicts…79-9343-00806d9d745a\scratchpad\envtest2026-10-26 11:39sonnet-5.569629,10729,00458,213$0.0829,10929,0591000
Worktree for parallel feature…79-9343-00806d9d745a\scratchpad\cp-only2026-10-26 11:40sonnet-5.5838132,83765,13498,360$0.0932,83932,6602000
Worktree for parallel feature…-4979-9343-00806d9d745a\scratchpad\both2026-10-26 11:44sonnet-5.5838015,45386,448102,289$0.0534,14933,9702000
7c5a8085…-4979-9343-00806d9d745a\scratchpad\both2026-10-26 11:45opus-5-5461717,54950,67968,849$0.1634,40734,1161000
Describe pending changesZ:\Sources\pkgtool2026-10-27 09:12opus-5-5283,33220,465561,428585,253$0.3455,29052,9027000
490e5fecZ:\Sources\pkgtool2026-10-27 09:13opus-5-5260214,34936,08351,036$0.1350,43450,4340110
d0ea1e0aZ:\Sources\pkgtool2026-10-27 09:15opus-5-5693716,074137,771154,788$0.1752,15951,2842110
Next steps after mergeZ:\Sources\pkgtool2026-10-27 09:16sonnet-5.5454036,563037,107$0.1036,56736,5671110
Branch comparisonZ:\Sources\pkgtool2026-10-27 10:13opus-5-5, opus-4-8, haiku-5-5, sonnet-5-5656277,0101,176,38753,230,44554,684,498$22.31285,378182,963371013
Package manifest validationZ:\Sources\usage-insights2026-10-27 14:02opus-5-5, haiku-5-5394155,017735,26426,145,54927,036,224$12.17232,450142,705261111
Plugin loading orderZ:\Sources2026-10-27 14:05opus-5-5164,46538,368462,415505,264$0.4972,70362,60011000
4596d48aC:\Users\jdoe2026-10-27 14:1100000$0.00002000
CLI argument parsingC:\Users\jdoe2026-10-27 14:41opus-5-51079616,417186,184203,407$0.1851,05150,6532000
Config file schemaC:\Users\jdoe2026-10-27 14:5000000$0.000014210
5d3a8d6b(no folder)2026-10-27 14:5200000$0.00009110
Error messages copy reviewZ:\Sources\Workshop2026-10-28 08:26opus-5-59414,03356,3883,027,8273,098,342$1.3490,76873,43639000
Docs site navigationZ:\Sources\Workshop2026-10-28 09:23opus-5-510451,277239,7305,864,4706,155,581$4.12168,655129,87959010
README quick startZ:\Sources\usage-insights2026-10-28 09:55opus-5-5296132,230612,24832,068,53032,813,304$13.96328,100220,818179010
Changelog automationZ:\Sources\Workshop2026-10-28 13:00opus-5-524837,507161,0217,413,3547,612,130$3.52121,40584,16276010
Code owners fileZ:\Sources\Workshop2026-10-28 13:24opus-5-56610,74575,6342,570,8412,657,286$1.33112,20598,02037010
Issue template updateZ:\Sources\Workshop2026-10-28 14:42opus-5-511445,88693,1844,765,2814,904,465$2.62130,19995,26651000
Total12,301,03711,605,60967,587,5733,204,956,6923,296,450,911$1,336.71785,048225,99729,645256259278
+

Installed plugins usage is per plugin, across its installs

+
+ + + + + + + + + +
PluginMarketplaceVersionScopeProjectEnabledUpdatedSkillsCmdsAgentsMCP callsHook runs
ccreportsccreports0.1.0useryes2026-10-2341000
datadogclaude-plugins-official0.7.14localZ:\Sources\AtlasRoot\atlas-workspace2026-06-2600000
frontend-designclaude-plugins-officialf359d43b4060projectZ:\Sources\atlas-workspaceno2026-10-1300000
frontend-designclaude-plugins-official1.1.1useryes2026-10-2800000
security-guidanceclaude-plugins-official2.0.17userno2026-10-2700000
skill-creatorclaude-plugins-officialf8c5e7a1dcb7useryes2026-10-2800000
superpowersclaude-plugins-official7.0.0userno2026-10-1600000
+

Adoption cohorts sessions by how they used Claude Code, GitHub Copilot and Cursor

+

Chat

Questions and exploration, no code changes

0.0 prs per session

Sessions
69
Share of sessions
35.0%
PRs
0
Edits per session
0.0
Median tool calls per prompt
2.0
Tokens per session
659,933
Cost per session
$0.44
+

Phase 1 · Assisted

The agent edits code, you steer step by step

0.6 prs per session

Sessions
32
Share of sessions
16.2%
PRs
18
Edits per session
16.2
Median tool calls per prompt
6.3
Tokens per session
6,880,530
Cost per session
$2.78
+

Phase 2 · Agent-first

One agent works through tasks on its own

0.5 prs per session

Sessions
47
Share of sessions
23.9%
PRs
23
Edits per session
21.2
Median tool calls per prompt
24.0
Tokens per session
13,750,578
Cost per session
$5.24
+

Phase 3 · Multi-agent

Subagents, background agents or worktrees

0.3 prs per session

Sessions
49
Share of sessions
24.9%
PRs
17
Edits per session
52.2
Median tool calls per prompt
48.5
Tokens per session
48,662,477
Cost per session
$19.82
+

Phase mix per week by the week a session started

+
+ + + + + + + + + + + + + + + + + + + + + + + + +
WeekSessionsChatPhase 1 · AssistedPhase 2 · Agent-firstPhase 3 · Multi-agentPRs
2026-02-02944.4%44.4%11.1%0.0%0
2026-02-16540.0%20.0%40.0%0.0%1
2026-03-09250.0%0.0%0.0%50.0%0
2026-03-161050.0%20.0%10.0%20.0%0
2026-03-23450.0%0.0%0.0%50.0%1
2026-04-06425.0%0.0%75.0%0.0%0
2026-04-13812.5%0.0%87.5%0.0%0
2026-04-2720.0%50.0%50.0%0.0%0
2026-05-0410.0%0.0%100.0%0.0%0
2026-05-11366.7%0.0%33.3%0.0%0
2026-05-25366.7%0.0%33.3%0.0%0
2026-06-0110.0%0.0%100.0%0.0%0
2026-06-1510.0%0.0%100.0%0.0%0
2026-06-2210.0%0.0%100.0%0.0%0
2026-08-312100.0%0.0%0.0%0.0%0
2026-09-142100.0%0.0%0.0%0.0%0
2026-09-21560.0%20.0%0.0%20.0%4
2026-09-281631.3%18.8%25.0%25.0%2
2026-10-053212.5%12.5%28.1%46.9%14
2026-10-122725.9%14.8%33.3%25.9%14
2026-10-193426.5%17.6%8.8%47.1%18
2026-10-262568.0%24.0%4.0%4.0%4
+

Agent adoption share of sessions using each capability

+
+ + + + + + + + + + + +
CapabilitySessionsShare of sessions
Code edits12865.0%
Pull requests4623.4%
Skills11055.8%
Subagents8241.6%
Background agents126.1%
Worktrees63.0%
Agent messaging31.5%
MCP tools3618.3%
Web search / fetch3618.3%
+

Usage by surface

+
+ + + + + + + + + +
SurfaceSessionsShare of sessionsPRs
CLI11156.3%56
Copilot in VS Code4824.4%2
Copilot CLI199.6%0
Agent SDK84.1%0
Cursor CLI73.6%0
Cursor editor21.0%0
Desktop app21.0%0
+

Claude Code recommendations most impact first; amounts at stake overlap

+
  1. 1

    Keep conversations short

    high impact · $668.50 at stake

    52% of your cost ($668.50) came from requests carrying more than 200k tokens of context; 16 sessions peaked above 400k (largest 785k).

    Start a new session, or /clear, when you switch tasks, and /compact with a hint of what to keep when a long task has to continue. Each request re-sends the whole context, so a turn at 400k tokens costs about ten times one at 40k.

  2. +
  3. 2

    Trim what Bash puts into the context

    high impact · $325.50 at stake

    Bash: 10,546 calls, ~858.5M tokens (28% of all tokens, ~$325.50), ~81k per call including re-reads.

    Make the repo's test and build scripts quiet (only failures), and ask in CLAUDE.md for short command output (head, grep, quiet flags). The bashOutputMaxChars setting caps what a command returns, and a PostToolUse hook can shorten a noisy command's output (updatedToolOutput). Every line of output stays in the context and is re-sent with each later request.

  4. +
  5. 3

    Run routine subagents on a cheaper model

    high impact · $238.66 at stake

    Subagents on claude-opus-5-5: general-purpose $356.93, Explore $32.18, critic $4.46, proptest $0.13. The same requests on Claude Sonnet 5.5 would have cost about $238.66 less.

    Set model: sonnet in the definition of your own agents that do routine work (.claude/agents/<name>.md), or model: haiku for search-only agents such as scouts. For built-in agents like general-purpose, set CLAUDE_CODE_SUBAGENT_MODEL (in the env of your settings), or ask Claude to delegate with a cheaper model (the Agent tool takes one per call). Keep the most capable model for agents that need deep reasoning.

  6. +
  7. 4

    Trim what Read puts into the context

    medium impact · $166.77 at stake

    Read: 2,411 calls, ~473.2M tokens (15% of all tokens, ~$166.77), ~196k per call including re-reads.

    Note in CLAUDE.md where things live, so Claude finds the right file without reading around it; split very large files; and ask for an Explore subagent for broad "find where…" questions, so the file contents stay out of your conversation. CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS (default 25,000) caps a single read.

  8. +
  9. 5

    Trim what Write puts into the context

    medium impact · $165.93 at stake

    Write: 832 calls, ~220.6M tokens (7% of all tokens, ~$165.93), ~265k per call including re-reads.

    Claude writes the whole file as output tokens, which then stay in the context: for large or repetitive files, ask for a script that generates them instead of the files themselves.

  10. +
  11. 6

    Avoid rebuilding the prompt cache

    medium impact · $82.30 at stake

    60 times a session re-read its context without a cache hit, costing $82.30 more than cache reads: 41 after a pause longer than the cache lifetime (avg 173k tokens reprocessed, $49.72); 12 without a pause, when something early in the conversation changed (avg 246k tokens reprocessed, $22.64); 7 after a model switch (avg 205k tokens reprocessed, $9.94).

    The main conversation's cache lasts 5 minutes with an API key or cloud provider and 1 hour on a Claude subscription; the promptCacheTtl setting (5m or 1h) chooses it, and 1-hour cache writes cost more. Before you continue a large conversation after a break, /compact it (or /clear for unrelated work) so a smaller context is rebuilt; /usage shows whether the cache is still warm. A miss without a pause means the start of the prompt changed. With tool search on (the default on supported models) MCP servers connecting mid-session no longer do this; with it off, servers connecting or tools being removed mid-session do, so keep the tool set stable during a conversation. Each model has its own cache: switching with /model mid-conversation re-reads the whole history. Pick the model at the start, or /compact before you switch.

  12. +
  13. 7

    Trim what Skill puts into the context

    low impact · $7.20 at stake

    Skill: 81 calls, ~23.6M tokens (1% of all tokens, ~$7.20), ~291k per call including re-reads.

    Keep your skills lean: put detailed instructions and examples in reference files the skill reads only when needed, since the whole SKILL.md is loaded into the context each time it runs.

  14. +
  15. 8

    Split long tasks before the context fills up

    low impact

    20 of 121 sessions hit compaction (24 times in total).

    Compaction replaces the conversation with a summary and loses detail. Break big work into tasks that fit one session, /clear between them, or run /compact yourself at a natural break with instructions on what to keep.

  16. +
  17. 9

    Run independent work in parallel

    low impact

    Background agents appear in 7% and worktrees in 5% of your sessions.

    Ask Claude to run independent tasks as parallel subagents: they run in the background while you keep working (Ctrl+B moves a long-running one there). When several agents change code at the same time, give each its own worktree: isolation: worktree in an agent's definition, or claude -w <name> for a whole session.

  18. +
  19. 10

    Fix or remove hooks that keep failing

    low impact

    6 hooks failed 2,181 times (463s spent): "$CLAUDE_PROJECT_DIR"/.claude/hooks/context_bundle_builder.py --type file_ops failed 935 of 935 runs (/usr/bin/env: 'uv run': No such file or directory); node .claude/hooks/protect-main.mjs failed 605 of 605 runs (Error: Cannot find module 'Z:\Sources\Nimbus\src\Nimbus.Agents\.claude\hooks\protect-main.mjs'); "$CLAUDE_PROJECT_DIR"/.claude/hooks/standing_brief_guard.py failed 248 of 248 runs (Python was not found; run without arguments to install from the Microsoft Store, or disable this shortcut from Settings > Apps > Advanced app settings > App ex…); ….

    A failing hook gives none of the protection it was added for. Use absolute paths ("$CLAUDE_PROJECT_DIR"/.claude/hooks/…) so hooks still resolve when Claude works in a subfolder, check that the interpreter (python, node) is on PATH, and remove stale ones from the settings file that defines them (/hooks lists them, read-only).

+

GitHub Copilot recommendations most impact first; amounts at stake overlap

+
  1. 1

    Delegate codebase exploration to subagents

    medium impact · $5.21 at stake

    Reading and searching files in your main conversation took ~137.7M tokens (65% of all), while 24% of your sessions launched a subagent (33 launches).

    Ask Copilot to use the explore agent for "find where…" and "how does … work" questions: it reads in its own context and returns a summary, so the file contents never stay in your main conversation.

  2. +
  3. 2

    Trim what view puts into the context

    medium impact · $4.34 at stake

    view: 2,416 calls, ~117.2M tokens (55% of all tokens, ~$4.34), ~48k per call including re-reads.

    Note in .github/copilot-instructions.md where things live, so Copilot opens the right file without reading around it; split very large files; and ask for the explore agent for broad "find where…" questions, so the file contents stay out of your conversation.

  4. +
  5. 3

    Trim what powershell puts into the context

    low impact · $1.01 at stake

    powershell: 1,334 calls, ~24.6M tokens (12% of all tokens, ~$1.01), ~18k per call including re-reads.

    Make the repo's test and build scripts quiet (only failures), and ask in .github/copilot-instructions.md (or AGENTS.md) for short command output (Select-Object -First, Select-String, quiet switches). Every line of output stays in the context and is re-sent with each later request.

  6. +
  7. 4

    Rein in mcp__github-mcp-server__get_job_logs's large results

    low impact · $0.00 at stake

    mcp__github-mcp-server__get_job_logs: 3 calls, ~1.1M tokens (1% of all tokens, ~$0.00), ~364k per call including re-reads.

    Its results are large: prefer tools or parameters that filter or page the results, turn off MCP servers you don't need for the task (/mcp), or ask Copilot to call it from a subagent so only a summary reaches your conversation.

  8. +
  9. 5

    Split long tasks before the context fills up

    low impact

    11 of 67 sessions hit compaction (14 times in total).

    Compaction replaces the conversation with a summary and loses detail. Break big work into tasks that fit one session, /clear between them, or run /compact yourself at a natural break.

  10. +
  11. 6

    Run independent work in parallel

    low impact

    Background agents appear in 4% of your sessions.

    Ask Copilot to run independent tasks as background agents: they keep working while the conversation goes on, and Copilot collects their results when they finish.

+
+ + +