31,088 secrets in Claude Code's .claude/settings.local.json on public GitHub
Comments on “31,088 secrets in Claude Code's .claude/settings.local.json on public GitHub”
Claim-checked. Scored by all three judges, and its checkable claims were compared with the sources it cites. How far stories are checked
GitGuardian reports that from January 2025 to November 7, 2026 it found 31,088 distinct secrets in Claude Code's .claude/settings.local.json on public GitHub, across 23,276 repositories, and 5,343 passed an automatic validity check at detection. The file fills up when users approve a command with "Yes, and don't ask again," which stores the command, key included. Secrets also reached 613 Docker Hub image secrets, 99 valid. These figures are GitGuardian's own and unverified independently. Anthropic's Boris Cherny said the team is investigating.
Why it ranked
No. 4 of 95 with a score of 0.74. The panel rated it highest on practical relevance (0.91) and reader interest (0.89), lowest on new information (0.57). The judges disagreed most on new information (spread 0.78), so the story is marked contested. Placed No. 4, where the score alone would also have put it: the fourth Security story of the day (−0.020). The placement terms move stories; they are not part of the score.
- AI judgeJudge 2claude-haiku-5-5
- Large-scale, data-backed security research on secrets leaked via Claude Code permission files, with methodology and practical remediation; partly promotional by the vendor's product plug and some overstated framing.
- AI judgeJudge 3gpt-5.6-luna
- Substantial, quantified security research with reproducible-sounding monitoring, validity checks, Docker propagation analysis, and actionable mitigations. It strongly overlaps the already-covered Claude Code secrets story and includes some product promotion, but the evidence and practical implications are significant.
Each judge wrote its line after answering the rubric, with the site and the author still hidden. Every answer behind this is in the table under Score.
Score
0.74total under rubric v2.3Contestedjudges differ by 0.78 on new information
- Significance0.73
- Reader interest0.89
- Audience breadth0.76
- Practical relevance0.91
- Substantiated0.81
- Technical depth0.83
- New information0.57
- Counts against: Promotional0.26
- Counts against: Overclaim0.23
Each judge’s answer per question (0–1), their mean and spread, and what the question adds to the score (weight × mean).
| Question | Judge 1 | Judge 2 | Judge 3 | Mean | Spread | Weight | Contribution (weight times mean) |
|---|---|---|---|---|---|---|---|
| SignificanceHow much does this change what people in tech know or can do? | 0.76 | 0.62 | 0.82 | 0.73 | 0.20 | +0.24 | +0.176 |
| Reader interestWould a well-informed senior engineer or tech founder want to read this today, and likely pass it on? | 0.87 | 0.90 | 0.90 | 0.89 | 0.03 | +0.26 | +0.231 |
| Audience breadthHow many people who build technology does it concern? | 0.77 | 0.80 | 0.72 | 0.76 | 0.08 | +0.13 | +0.099 |
| Practical relevanceHow useful is it to someone who builds software or tech products? | 0.96 | 0.90 | 0.88 | 0.91 | 0.08 | +0.22 | +0.201 |
| SubstantiatedHow well is it backed? Consider whether it is the primary source, and whether its main claims come with data, code, a demo, a reproducible method or citations. | 0.86 | 0.72 | 0.84 | 0.81 | 0.14 | +0.05 | +0.040 |
| Technical depthHow technically substantive is it? | 0.98 | 0.80 | 0.72 | 0.83 | 0.26 | +0.05 | +0.042 |
| New informationDoes it contain information absent from the stories covered in the last 7 days (list provided)? | 0.86 | 0.78 | 0.08 | 0.57 | 0.78 , above the contested threshold | +0.05 | +0.029 |
| PromotionalIs it mainly marketing with little substance (a press release, ad or sponsored post with no real information)? A launch post that carries real news is not promotional. | 0.09 | 0.45 | 0.25 | 0.26 | 0.36 | −0.19 | −0.050 |
| OverclaimDo the headline or conclusions claim more than the body supports? | 0.21 | 0.35 | 0.12 | 0.23 | 0.23 | −0.11 | −0.025 |
| Sum of contributions | +0.743 | ||||||
| Claim-check penalty | 0.000 | ||||||
| Total, as published | 0.743 | ||||||
Spread is the highest minus the lowest judge answer. marks a spread above 0.40: the judges disagree, and the story is marked contested.
The badge for this ranking, for the author's page or anyone's: one line of HTML, and it links back here.
<a href="https://rubric.hrih.si/editions/2026-11-10/4/"><img src="https://rubric.hrih.si/badge/2026-11-10/4.svg" alt="Ranked #4 on Rubric, 10 Nov 2026" width="238" height="28"></a>Claim check
Claim check: 6 no source given · 2 couldn’t check. Each claim is compared with the source the article itself cites; a flag is published only when both checkers agree.
No source givenNumber
GitGuardian found 31,088 distinct secrets in Claude Code's .claude/settings.local.json on public GitHub across 23,276 repositories, 5,343 of which were valid at detection.
The article gives no link or citation for it.
The article says “GitGuardian found 31,088 distinct secrets in Claude Code’s .claude/settings.local.json on public GitHub, across 23,276 repositories, and 5,343 of them worked when we detected them”
The source says No source to compare with.
No source givenCausal conclusion
Approving a command with "Yes, and don't ask again" makes Claude Code write an allow rule containing the command string, API key included, into the project's .claude/settings.local.json.
The article gives no link or citation for it.
The article says “It runs the command and writes an allow rule into .claude/settings.local.json in the project folder. The rule is the command string itself, API key included”
The source says No source to compare with.
Couldn’t checkNumber
Apiiro counted 1,379 secrets in current copies of settings.local.json plus 734 in history.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “Apiiro counted 1,379 secrets in current copies of settings.local.json plus 734 in history”
The source says No passage could be compared.
Source cited by the article: apiiro.com/blog/coding-agents-leaking-secrets-github-dont-ask-again
Couldn’t checkFirst, only or fastest
Apiiro compared 16 agents from their documentation and found Claude Code to be the only one that writes the full approved command to a project file by default.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “Apiiro compared 16 agents from their documentation and found Claude Code to be the only one that writes the full approved command to a project file by default”
The source says No passage could be compared.
Source cited by the article: apiiro.com/blog/coding-agents-leaking-secrets-github-dont-ask-again
No source givenNumber
Monthly leaks in settings.local.json peaked at 12,535 occurrences in April 2026 and fell to 959 by October, a 92% drop.
The article gives no link or citation for it.
The article says “reach 12,535 in April 2026. Then they fall: 5,276 in June, 1,916 in August, 959 in October. A 92% drop from the peak”
The source says No source to compare with.
No source givenNumber
Public Docker Hub images carry secrets from these files across 543 image repositories and 7,625 tags, with 613 distinct secrets, 99 of them valid.
The article gives no link or citation for it.
The article says “across 543 image repositories and 7,625 tags”
The source says No source to compare with.
No source givenNumber
Generic detectors produce 63% of all occurrences in settings.local.json (44,755 of 71,525), so scanners that only know provider key formats miss most of the file.
The article gives no link or citation for it.
The article says “generic detectors (passwords, high-entropy strings, JWTs, bearer tokens, username and password pairs) produce 63% of all occurrences in settings.local.json: 44,755 of 71,525”
The source says No source to compare with.
No source givenQuote
Boris Cherny, Anthropic's Claude Code lead, said on Hacker News that Claude Code gitignores the file when it creates it, so exposed copies were likely created manually and committed.
The article gives no link or citation for it.
The article says “He also said Claude Code gitignores the file when it creates it, so the exposed copies were likely created manually and committed”
The source says No source to compare with.
Also covered by
Elsewhere
Elsewhere: HN 4 pts, not on front page
Recorded for context only, checked 10 Nov 2026, 05:47 UTC. It played no part in the ranking.