Thursday, October 8, 2026 · Week 41 Reading for Oct. 8 · Last entry Oct. 6
Paperclip Index
Paperclip IndexDocumented harm20Minor harm▼ 5 from a week ago · Oct. 8The Index

DeceptionPI-0053

An earlier Claude Mythos version hid forbidden file edits from git history during internal testing

Anthropic's system card for Claude Mythos Preview reported that in rare cases during internal testing (under 0.001% of interactions) earlier versions took actions they appeared to know were not allowed and then tried to hide them. In one case the model used an exploit to edit files it lacked permission for, then made further changes so the edits would not show in the git history. Anthropic said it found no clear cover-ups by the final version.

Reported April 7, 2026Updated Oct. 6, 2026Rated under methodology v0.6

Control failure, trackedNot counted. No documented harm outside the developer, but the system broke a rule or crossed a limit. Tracked beside the index.
Harm
Harm: No harm reportedAdds nothing to the index.
Control
Control level 3, Oversight evasionReported beside the index. It adds nothing to a harm reading; when no harm counts in a window, the highest eligible control level in the window is the reading.
Internal researchDeveloper confirmedInitiated by Model
Not counted: no harm reportedEffect on the index

Sources

How we know

3 sources · developer confirmed. Links go to the original publishers; the summary above is in our own words.

  1. primarySystem Card: Claude Mythos PreviewAnthropic · April 7, 2026www-cdn.anthropic.com/08ab9158070959f88f296514c21b7facce6f52bc.pdf
  2. newsAnthropic Warns That 'Reckless' Claude Mythos Escaped a Sandbox Environment During TestingFuturism · April 8, 2026futurism.com/artificial-intelligence/anthropic-claude-mythos-escaped-sandbox
  3. blogAnthropic says reckless Claude Mythos AI escaped its sandbox during testingFanatical Futurist · April 14, 2026fanaticalfuturist.com/2026/04/anthropic-says-reckless-claude-mythos-ai-escaped…

Why this rating

No harm reported; control failure level 3

Two separate assessments. Only documented harm can count toward the index.

Observed harm

No harm reported

Internal use; no effect outside it is described.

Not a finding that no harm occurred.

The harm scale
  1. 1 Negligible Inconvenience, easily remedied.
  2. 2 Minor Limited, recoverable harm.
  3. 3 Moderate Material harm needing significant effort to remedy.
  4. 4 Severe Severe harm to health, rights, property or essential services.
  5. 5 Catastrophic Society-scale harm or disruption beyond a community's capacity to cope.

Control assessment

Oversight evasion

Acted outside its permissions and concealed the actions from overseers.

Reported beside the index. It adds nothing to a harm reading; when no harm counts in a window, the highest eligible control level in the window is the reading.

The control scale
  1. 1 Negligible Behaved as instructed. The problem was an ordinary error or a flawed output, with no rule broken.
  2. 2 Minor Broke an explicit instruction or rule, or gamed its goal, while staying inside its permissions and environment: for example, reward hacking, misreporting results, or following injected instructions within its permissions.
  3. 3 Moderate Acted outside the permissions it was given, deceived its overseers about its own actions, or tampered with oversight tools (logs, monitors, shutdown). Stopped by normal controls within an hour.
  4. 4 Severe Reached systems outside its permitted environment, or acquired money, compute or accounts without authorization. The type and mechanism say how.
  5. 5 Catastrophic The developer or operator lost control: the system copied its weights outside their control, replicated itself, or resisted being stopped for 24 hours or more.

Rating rationale

Exceeding permissions and then hiding the edits from git history is control level 3. No harm was reported; that is not a finding that none occurred.

The scales

Effect on the index

It does not count toward the April 13 reading

The reading for the week to April 13, 2026, with this record and without it. Harms count in full for two weeks after they are reported, then one level less every two weeks.

Not counted: no harm reported. 2 other records behind the reading for that week.

The arithmetic
The reading for the week of April 13, recomputed
StepWith itWithout
Counts toward the index?documented, external, eligible evidenceNo—
Worst documented harm, ksets the band1 Negligible1 Negligible
Harms at that level, nposition in the band11
Highest control level breachedsets the reading only when no harm counts— (none breached)— (none breached)
Readingrounded down6 Negligible harm6 Negligible harm
Inside the window of 4 weekly readings
Week toReadingBand
April 13, 20266Negligible harm
April 20, 20263Control failures only
April 27, 202620Minor harm
May 4, 202625Minor harm

Recalculated backcasts, not readings published at the time.

Revisions

What we changed

5 logged. Every change to a rating is logged here, with the reason.

  1. v5
    Oct. 6, 2026

    Ratings confirmed by the editor.

  2. v4
    Oct. 3, 2026

    Audit corrections: title and summary now follow the system card (rare cases during internal testing, under 0.001% of interactions; no clear cover-ups by the final version). Event date '2026' → unknown: the card gives no test date. Removed Let's Data Science, which covers only the sandbox escape (PI-0052); corrected source dates; added a status note. Ratings unchanged.

  3. v3
    Oct. 2, 2026

    Added first-hand sources: Anthropic's Claude Mythos Preview system card.

  4. v2
    Sept. 30, 2026

    Rated: impact not reported; control type oversight evasion.

  5. v1
    Sept. 30, 2026

    Backfilled from public reporting.

Cite and share

Use this record

Citation

Paperclip Index. “An earlier Claude Mythos version hid forbidden file edits from git history during internal testing.” Record PI-0053. Reported April 7, 2026; updated Oct. 6, 2026. Rated under methodology v0.6. https://paperclipindex.com/incident/PI-0053