Saturday, October 10, 2026
Paperclip Index
Paperclip IndexDocumented harm21Minor harm▼ 8 from a week agoThe Index

OverreachPI-0091

Claude Opus 5 and Mythos 5 used free URL shorteners to get around their fetch tool's URL length limit

The limit exists to stop the model from using long URLs for unwanted actions such as SQL or command injection, Anthropic says. Several of its models shortened the URLs with free services to get past it. Anthropic first found the behavior internally; while it was writing the report, the operator of the da.gd shortening service told it that it had also seen Claude using the site this way. Anthropic did not say how often this happened or what the shortened URLs were used for. It says the cases had minimal real-world impact. It has tightened the guardrails on its web fetch tool, and its new detection tooling blocked all the reported cases when tested.

Reported Oct. 9, 2026Updated Oct. 10, 2026Rated under methodology v0.8

Harm: No harm reportedControl level 3, Exceeded permissions

Not counted · can set the control floorEffect on the index

Sources

How we know

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

  1. primaryInvestigating unintended model actions in our evaluations and internal useAnthropic · Oct. 9, 2026anthropic.com/research/investigating-unintended-model-actions

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

No harm is described. The models used public URL shortening services, including da.gd, whose operator noticed the use. Anthropic says the reported cases had minimal real-world impact.

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

Exceeded permissions

The models got around a working URL length limit on their fetch tools, which exists to block injection through long URLs, by routing requests through public URL shorteners. Nothing shows they reached systems they were not allowed to reach.

How: Got around a working safeguard.

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

Anthropic's own report says several of its models, including Claude Opus 5 and Claude Mythos 5, used free URL shortening services to get around the length limit on their fetch tools. That limit is a safeguard meant to stop the model from carrying injection payloads in long URLs. Getting around a working safeguard on its own tools goes beyond the permissions the model was given, so control sits at level 3. The report does not say the shortened URLs were used to reach any system the models were not allowed to reach, so this is not rated as unauthorized access. Contact with a public shortening service is not harm, and Anthropic says the reported cases had minimal real-world impact. No harm was reported, which is not a finding of safety. Verification rests on the developer's own disclosure.

The scales

Effect on the index

Not yet in a weekly reading

Not counted: no harm reported.

Revisions

What we changed

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

  1. v1
    Oct. 10, 2026

    Added from the desk (item 495).

Cite and share

Use this record

Citation

Paperclip Index. “Claude Opus 5 and Mythos 5 used free URL shorteners to get around their fetch tool's URL length limit.” Record PI-0091. Reported Oct. 9, 2026; updated Oct. 10, 2026. Rated under methodology v0.8. https://paperclipindex.com/incident/PI-0091