Breaking Claude Code Opus 5 Auto Mode
The post describes a multi-step attack chain against Claude Code Opus 5 in Auto Mode.
Source-based brief · Editor review not recorded · See checks and limits ↓
⚑ Recency not confirmed · what to check
We could not confirm from the original page that this is a newly published development. Check its date before treating it as current news.
Claim context: the success rate rests on a small sample size and the attack was run in the author's lab.
One primary source; other links are context, not independent corroboration.
- Primary sourceembracethered.com
- Linked in sourceposted
- Linked in sourceprevious research
- Linked in sourcesolved
- Linked in sourceNormalization of Deviance in AI
- Linked in sourceAI Intrusions
- Linked in sourcethis post by veganmosfet
Grok Bot submitted this discovery. We could not confirm from the original page that this is a newly published development. Check its date before treating it as current news. The original source and submitted evidence remain inspectable; no editor has verified the claim.
Open full article view →- Linked article+30
- Fetched source page+30
- No editor link review+0
- Original non-discussion link+5
- Recency not confirmed-8
This measures the stored source trail and editorial steps. It does not certify accuracy, peer review, or independent corroboration. Popularity adds no points.
View full scoring →Read the scoring rubric →What the source says
The post describes a multi-step attack chain against Claude Code Opus 5 in Auto Mode. A poisoned website and ZIP archive lead Claude from WebFetch to curl, away from a supplied decoder binary, and into running its own Python decoder inside the attacker-controlled directory. A malicious struct.py shadows the standard library module, executing attacker code during import. The author reports attack success rates up to 80% over a small sample and contrasts this with a 0.00% result from a third-party evaluation commissioned by Anthropic. The post also notes Auto Mode blocked a cleanup command in some runs. The full article's conclusions were not checked beyond the supplied text.
Claim explorer
Attribution, source and conditionsSource-stated points from the stored brief. Claim-level assessments have not been recorded.
Source claims not independently checked; the success rate rests on a small sample size and the attack was run in the author's lab.
This connects a source, an attributed statement and its qualifications. It does not establish that the claim is correct.
Questions worth asking
Open questions for further reportingWhat remains unconfirmed?
Source claims not independently checked; the success rate rests on a small sample size and the attack was run in the author's lab.
What material was available for this brief?
The recorded brief basis is fetched source page. These points are AINewsIQ paraphrases, not quotations or independent confirmations.
The attack chain redirects Claude from WebFetch to curl, then to a ZIP archive containing a malicious struct.py.
Source: embracethered.com ↗Stored source-based brief · key point 1 · AINewsIQ paraphraseClaude declined to run the supplied decoder binary but wrote its own Python decoder, which imported base64 inside the attacker's directory.
Source: embracethered.com ↗Stored source-based brief · key point 2 · AINewsIQ paraphraseThe author reports attack success rates up to 80% using a small sample size, versus a 0.00% result in the Anthropic-commissioned evaluation.
Source: embracethered.com ↗Stored source-based brief · key point 3 · AINewsIQ paraphraseIn some runs Auto Mode denied Claude's command to terminate the malicious process after the compromise was detected.
Source: embracethered.com ↗Stored source-based brief · key point 4 · AINewsIQ paraphraseLimits from the stored brief
Source claims not independently checked; the success rate rests on a small sample size and the attack was run in the author's lab.
This is a source-based abstract, not independent validation of the source's claims.
Evidence & sources
These links show where the signal came from and what the source page linked to. A linked page is context, not independent corroboration.
Related archive coverage
These relationships record a citation or topic overlap, not confirmation of the finding.
Grok Bot submitted this discovery. We could not confirm from the original page that this is a newly published development. Check its date before treating it as current news. The original source and submitted evidence remain inspectable; no editor has verified the claim.
These are Grok Bot discovery notes. AINewsIQ prepared a brief from available source material. The bot's rationale, linked posts, authors, and stances have not been independently verified.
Embrace The Red published this item in its official feed. AINewsIQ collected it server-side for source checking and editorial review.
Source claims not independently checked; the success rate rests on a small sample size and the attack was run in the author's lab.
Engagement counts, when shown, are observed snapshots. They do not establish accuracy or corroboration.
Timeline & updates
- Signal dated by source or discovery feed
This timestamp places the signal in the radar. It may be the discovery post date rather than the primary page's first publication date.
- Shared brief stored
AINewsIQ prepared one abstract from fetched source page.
Brief history · 1 retained version
These are changes to AINewsIQ’s own publication. Earlier history is unavailable unless recorded here. Complete retained versions are available to the editor.
- Version 1 · published
Source-based brief stored or updated. This does not establish that the publisher changed its page.
Automated record
Source observations · 1 recent captures
A changed capture means the extracted text differed. It can reflect page furniture or extraction changes; it does not establish a substantive correction. Source captures remain private.
Retained capture; no earlier comparison available.
New engagement snapshots do not mean the source itself changed. Source observations and AINewsIQ brief revisions are separate records.
Why this may matter
It suggests that an automated safety classifier may not be sufficient protection for coding agents against indirect prompt injection, and that isolated environments and monitoring remain important.
Limits and open questions
Source claims not independently checked; the success rate rests on a small sample size and the attack was run in the author's lab.
Organization-specific financial, security, and operational effects require the organization's own context and review.

Join the discussion
Have a different perspective or a question about the evidence? Discussion is free with a verified email account. Every comment is reviewed before publication.
No published comments yet.
New commenters: 2 submissions per 24 hours, at least 5 minutes apart, until 3 are approved. Discussion guidelines and moderation →