codex usage

    Codex Weekly Usage Report — September 7–13, 2026

    Executive summary

    Rock Paper Scissors review and operations led the observed September 7–13 week.

    Available Skysight evidence across five days contains 60 active public summaries across six work arcs, led by Rock Paper Scissors operations. Review and quality assurance is the strongest work-mode signal, followed by operations and fulfillment. The record includes pick-list preparation and broader delivery and review signals; three follow-up categories remain open. September 9 and 12 have no available Skysight evidence, so their activity remains unknown.

    Codex execution telemetry is supporting evidence, not a measure of productivity, duration, or the week's value.

    6
    Work arcs

    Distinct public work threads supported by retained Skysight summaries.

    3
    Outcome signals

    Public-safe outcomes or decisions supported by observed context.

    3
    Open follow-ups

    Follow-ups the observed record does not show as resolved.

    29
    Top-level tasks

    Top-level Codex tasks recorded in local execution logs; not a productivity score.

    Week in reviewObserved Sep 7, 6:40 AM–Sep 13, 3:30 PM

    What shaped the week

    Major work arcs

    Rock Paper Scissors operations and Other observed work are the two largest observed arcs by retained summary mentions. Counts can overlap and do not represent duration.

    Observed evidence by work arc

    Retained structured-summary mentions; overlapping evidence, not time allocation.

    Rock Paper Scissors operations
    33
    Other observed work
    17
    Hyphenomenon product and reporting
    5
    Product strategy
    5
    Research and administration
    4
    Mailbox and customer operations
    2

    Rock Paper Scissors operations

    33 observed

    Catalog quality review and wholesale or fulfillment preparation.

    Other observed work

    17 observed

    Observed work that did not match a named public reporting arc.

    Hyphenomenon product and reporting

    5 observed

    Product-system review, reporting strategy, and project sequencing.

    3 additional work arcsLower-volume observed threads remain available without competing with the primary arcs.

    Product strategy

    5 observed

    Product framing, research, and strategic decision-making.

    Research and administration

    4 observed

    Research, purchasing, and administrative follow-through.

    Mailbox and customer operations

    2 observed

    Mailbox, customer-support, and operational follow-through.

    Outcomes and decisions

    • Faire fulfillment input prepared

      Advanced

      Prepared the pick-list input for a size-counting and fulfillment task.

    • Delivery and implementation progress

      Advanced

      Retained summaries indicate implementation, publication, deployment, or other delivery progress.

    • Review and verification decision

      Advanced

      Retained summaries indicate an approval, acceptance, verification, or passed review gate.

    Next steps and open loops

    • Pick-list size totals

      Open

      Complete and verify the size totals requested from the Faire pick list.

    • Outstanding follow-through

      Open

      Retained summaries identify work that remained pending or still needed follow-through.

    • Unresolved quality or reliability issue

      Open

      Retained summaries identify an unresolved issue, defect, failure, or remediation need.

    How the observed week unfolded5 day summaries
    1. Sep 7

      Rock Paper Scissors operations + Other observed work

      20 active summaries, led by review and quality assurance signals. 1 additional observed summary was protected.

    2. Sep 8

      Rock Paper Scissors operations + Other observed work

      8 active summaries, led by operations and fulfillment signals.

    3. Sep 10

      Other observed work + Product strategy

      9 active summaries, led by review and quality assurance signals. 2 additional observed summaries were protected.

    4. Sep 11

      Rock Paper Scissors operations + Hyphenomenon product and reporting

      20 active summaries, led by review and quality assurance signals.

    5. Sep 13

      Other observed work + Rock Paper Scissors operations

      3 active summaries, led by research and administration signals.

    How to read this evaluationWork-mode signals, protected context, and evidence limits

    Observed work-mode signals

    Overlapping summary-level signals, not duration or a complete time sheet.

    Review and quality assurance
    25
    Operations and fulfillment
    20
    Building and delivery
    10
    Research and administration
    8
    Planning and decisions
    5

    Privacy and evidence boundary

    3 observed summaries were excluded from public topic analysis. Their time positions remain visible so the report does not mislabel protected context as missing or inactive.

    This evaluation uses 75 summaries from five isolated Skysight source-contract v2 remote restores, each hash-verified with zero conflicts or disappeared records. September 9 and 12 are missing and unverified. Three protected positions remain visible without exposing topics; capture completeness is not asserted. Counts are overlapping evidence signals rather than duration, productivity, value, or unique task totals. Uncovered intervals remain unknown rather than zero activity; retained local Codex telemetry excludes other-device or cloud-only activity absent from those logs.

    Evidence appendix

    Repository activity, delivery records, source evidence, and technical diagnostics support the evaluation without competing with its first read.

    Explore supporting evidenceExecution, delivery, Skysight source context, and technical telemetry

    Where The Work Landed

    Repository activity and changed lines show which codebases received attention during the report window.

    Tool Calls By Repository

    Working repository resolved from tool workdir or turn context.

    shopify-headless916 43%
    multi-mailbox-ops892 42%
    hyphenomenon234 11%
    chronicle-visualizer66 3%
    Outside git metadata13 <1%
    aerial-platform3 <1%
    self-hosted-ai-starter-kit1 <1%

    Verified Delivery Ledger

    Outcomes come from terminal return records. Intent-only command counts remain in the technical appendix.

    Command Outcomes

    Pull request operations - unknown
    Count38
    Share62%
    Commit operations - success
    Count8
    Share13%
    Commit operations - unknown
    Count5
    Share8%
    Content push operations - success
    Count5
    Share8%
    Pull request operations - success
    Count3
    Share5%
    Content push operations - unknown
    Count2
    Share3%

    Repositories In Successful Push Output

    shopify-headless
    Count1
    Share100%

    Agent Stack Used

    Observed runtime evidence only; the available catalog is excluded. Status chips distinguish completed calls, skill loads, and explicit requests.

    App or connectorNo observed call
    Runtime providerMulti Mailbox Ops
    Loaded skillmulti-mailbox-ops
    • Invoked means a completed runtime call was recorded.
    • Loaded means a skill file was read or discovered.
    • Requested marks an explicit skill request in the source.
    • The logs do not preserve a reliable causal chain between every layer, so the diagram shows observed layers rather than inventing one.

    Tool Providers

    Completed MCP provider calls, grouped by server namespace.

    Multi Mailbox Ops
    Invoked1 rollout
    102

    Skills

    Explicitly requested or loaded skills; application is not inferred.

    multi-mailbox-ops
    Loaded3 rollouts
    76
    autonomous-backlog-runner
    Loaded7 rollouts
    9
    shopify-admin
    Loaded4 rollouts
    9
    control-in-app-browser
    Loaded3 rollouts
    8
    visual-design-critique
    Requested7 rollouts
    8
    shopify-admin-execution
    Loaded3 rollouts
    6
    ci-hosting-manual-gates
    Loaded3 rollouts
    3
    rps-etsy-shop-uploader-listing-workflow
    Loaded1 rollout
    3
    control-chrome
    Loaded1 rollout
    2
    frontend-app-builder
    Loaded1 rollout
    2

    Skysight Evidence Across The Week

    This evidence view locates the public work arcs and protected-context positions that support the synthesis. The calendar grid is not expected computer activity, and values are not duration.

    12–4a4–8a8a–12p12–4p4–8p8p–12a

    Rock Paper Scissors operations

    Other observed work

    Hyphenomenon product and reporting

    Product strategy

    Research and administration

    Mailbox and customer operations

    Protected context

    Color intensity represents overlapping structured-summary mentions. Empty outlined cells have no retained summary evidence; zero means retained evidence exists but this domain did not match. Values are not duration.
    Work domain12–4a4–8a8a–12p12–4p4–8p8p–12a
    Rock Paper Scissors operationsRock Paper Scissors operations, 12–4a: no retained summary evidenceRock Paper Scissors operations, 4–8a: 10 summary mentionsRock Paper Scissors operations, 8a–12p: 14 summary mentionsRock Paper Scissors operations, 12–4p: 7 summary mentionsRock Paper Scissors operations, 4–8p: 2 summary mentionsRock Paper Scissors operations, 8p–12a: no retained summary evidence
    Other observed workOther observed work, 12–4a: no retained summary evidenceOther observed work, 4–8a: 2 summary mentionsOther observed work, 8a–12p: 9 summary mentionsOther observed work, 12–4p: 5 summary mentionsOther observed work, 4–8p: 1 summary mentionOther observed work, 8p–12a: no retained summary evidence
    Hyphenomenon product and reportingHyphenomenon product and reporting, 12–4a: no retained summary evidenceHyphenomenon product and reporting, 4–8a: 1 summary mentionHyphenomenon product and reporting, 8a–12p: 2 summary mentionsHyphenomenon product and reporting, 12–4p: 2 summary mentionsHyphenomenon product and reporting, 4–8p: zero matches within retained summariesHyphenomenon product and reporting, 8p–12a: no retained summary evidence
    Product strategyProduct strategy, 12–4a: no retained summary evidenceProduct strategy, 4–8a: 4 summary mentionsProduct strategy, 8a–12p: zero matches within retained summariesProduct strategy, 12–4p: zero matches within retained summariesProduct strategy, 4–8p: 1 summary mentionProduct strategy, 8p–12a: no retained summary evidence
    Research and administrationResearch and administration, 12–4a: no retained summary evidenceResearch and administration, 4–8a: 3 summary mentionsResearch and administration, 8a–12p: 1 summary mentionResearch and administration, 12–4p: zero matches within retained summariesResearch and administration, 4–8p: zero matches within retained summariesResearch and administration, 8p–12a: no retained summary evidence
    Mailbox and customer operationsMailbox and customer operations, 12–4a: no retained summary evidenceMailbox and customer operations, 4–8a: zero matches within retained summariesMailbox and customer operations, 8a–12p: zero matches within retained summariesMailbox and customer operations, 12–4p: 2 summary mentionsMailbox and customer operations, 4–8p: zero matches within retained summariesMailbox and customer operations, 8p–12a: no retained summary evidence
    Protected contextProtected context, 12–4a: no retained summary evidenceProtected context, 4–8a: 1 withheld summary mentionProtected context, 8a–12p: 2 withheld summary mentionsProtected context, 12–4p: zero matches within retained summariesProtected context, 4–8p: zero matches within retained summariesProtected context, 8p–12a: no retained summary evidence

    Structured Evidence Coverage

    Retained summaries
    75
    Active public summaries
    60
    Protected summaries
    Topic excluded; observed position retained.
    3
    Inactive or locked summaries
    12
    Covered intervals
    75
    Expected calendar intervals
    Calendar grid only; not expected computer activity.
    1,008

    Codex Execution Activity

    Token and hourly activity describe the agent execution layer. They are supporting diagnostics, not a measure of the week's value, duration, or human productivity.

    Token Usage

    Cumulative runtime counters are differenced per rollout. Repeated snapshots are ignored; cached input stays visible instead of being mistaken for fresh context.

    Token Usage

    Retained local Codex telemetry only; not an account-wide usage or billing statement.

    317,703,564
    Total processed
    Input + output
    Input
    316,684,776
    Cached input
    303,777,152
    Non-cached input
    12,907,624
    Output
    1,018,788
    Reasoning output
    243,625

    Cached input is 96% of input. Reasoning output is included within output tokens.

    Tokens By Day

    Processed input plus output assigned to the local day of each positive cumulative-counter increment.

    Mon Sep 7108,959,953 34%
    Thu Sep 1065,899,751 21%
    Fri Sep 1164,630,901 20%
    Tue Sep 838,835,112 12%
    Sun Sep 1320,172,294 6%
    Sat Sep 1219,205,553 6%

    Activity By Day And Clock Hour

    The day row shows each date in the report window. The hour chart below combines the same clock hour across all seven days; it is not a single-day timeline.

    Daily session timeline
    1. Mon Sep 78
    2. Tue Sep 84
    3. Wed Sep 90
    4. Thu Sep 105
    5. Fri Sep 113
    6. Sat Sep 123
    7. Sun Sep 134

    Each count is a de-duplicated session start assigned to that local calendar day.

    Clock-hour pattern across the full week
    12 AM
    37
    2 AM
    2
    4
    4 AM
    64
    280
    5c
    6 AM
    200
    114
    8 AM
    553
    195
    10 AM
    65
    80
    12 PM
    48
    124
    2c
    2 PM
    82
    1c
    40
    4 PM
    53
    45
    6 PM
    69
    21
    8 PM
    22
    21
    10 PM
    6

    Total: 2,125 tool calls. “c” marks successful commit operations in that hour.

    Technical Telemetry

    Raw tool, event, workspace, and patch diagnostics used to audit the analyzer. These counts are intentionally secondary to outcomes.

    Tool Surfaces

    Terminal/local shell2,082 98%
    Long-running task waits23 1%
    Other/uncategorized15 <1%
    Node/browser automation5 <1%

    Function Calls

    wait23 53%
    request_user_input_async7 16%
    sleep6 14%
    js5 12%
    request_user_input2 5%

    Custom Tool Calls

    Specific custom calls and their share of all recorded tool events.

    exec
    Count2,082
    Share98%

    Response Events

    event_msg
    Count9,018
    Share49%
    response_item
    Count6,877
    Share37%
    token_usage_record
    Count2,281
    Share12%
    turn_context
    Count155
    Share<1%
    world_state
    Count89
    Share<1%
    session_meta
    Count27
    Share<1%
    compacted
    Count22
    Share<1%

    Output Records

    Tool output records
    Count2,236
    Share77%
    Git refs in output
    Count477
    Share16%
    Truncated outputs
    Count161
    Share6%
    GitHub PR links emitted
    Count41
    Share1%
    Long-running command handoffs
    Count1
    Share<1%

    Session And Log Audit

    Active rollout log files
    Count29
    Share35%
    Raw session_meta records
    Count27
    Share33%
    Unique session_meta.session_id values
    Count27
    Share33%

    Token Counter Audit

    Metered model-response increments
    2,259
    Token-count records
    2,301
    Repeated snapshots ignored
    42
    Records missing cumulative usage
    0
    Cumulative counter resets
    8

    Source Coverage

    Configured local session stores
    2
    Candidate rollout files inspected
    46
    Active rollout files
    29