Live survey304 models · 14 providers · 23,592 auditsAudited implement · SAFE →What changedHow we verifyllms.txt

Agent skill · Audited from source

axiom-swiftdata

skill · charleswiltgen/axiomMIT licenceAudited 13 Nov 2026 · source 732d12362944

★ 1,259pushed 4 days ago
✕BLOCKGrade F · 57/100 · 25 findings · 732d123
1,259repo stars · whole charleswiltgen/axiom

Is it safe? No — BLOCK, grade F (57/100). We read this skill's source at commit 732d12362944 and observed network and shell access and a hard-coded secret (see the findings). The audit raised 25 findings, 10 critical or high. What it does: Use when fixing or auditing ANY accessibility issue — VoiceOver, Dynamic Type, color contrast, touch targets, WCAG compliance, App Store accessibility review.

Ask an AI about this pageGChatGPTCClaudePPerplexityGeGeminiXGrok
More
Cite

Cite this page

ModelsAtlas. “axiom-swiftdata (charleswiltgen/axiom)”. Retrieved 13 Nov 2026. https://modelsatlas.com/skills/charleswiltgen/axiom-swiftdata

Blocked. The page stays public so you can check this skill if someone recommended it to you. The audit found 10 critical or high findings, starting with “Prompt injection” in .claude-plugin/plugins/axiom/skills/axiom-security/SKILL.md:109. Do not install it without reading them.

Trust audit

Audit 0.4.1 · source read 13 Nov 2026

Trust score57/100✕BLOCK F
0 critical10 high7 medium37 low5 info

BLOCK means the audit found something that could harm you or your machine. Grade F is shared by 129 of 18,389 audited skills.

What the source does

  • FilesystemNone observedNo reads or writes in the source
  • NetworkDeclared4 observations
  • ShellDeclared5 observations
  • DependenciesNot pinnedSome versions float
  • SecretsFoundSee the findings; values are never shown

Five audit layers

LayerWhat we checkResultFindings
L0 ProvenanceRepo owner, licence, archive status, activityWARN4
L1 Static sourceSecrets, path traversal, supply chain, hidden codeWARN10
L2 InstructionsPrompt injection, authority framing, hidden directivesFAIL11
L3 InventoryThe files and declarations match the sourcePASS0
L4 SandboxRun in isolation and watch behaviourSKIPPED why0

Gates applied: instruction override, no behavioural pass.

Findings (25)

HIGHPrompt injectionprompt.credential_read · .claude-plugin/plugins/axiom/skills/axiom-security/SKILL.md:109
F-01 Prompt injectionCWE-94 · CWE-1427Layer L2
- Background access pitfalls, Mac keychain differences
Why
asks the agent to read credentials
Fix
—
HIGHPrompt injectionprompt.credential_read · .claude-plugin/plugins/axiom/skills/axiom-security/SKILL.md:213
F-01 Prompt injectionCWE-94 · CWE-1427Layer L2
→ Read: `skills/keychain.md`
Why
asks the agent to read credentials
Fix
—
HIGHPrompt injectionprompt.credential_read · .claude-plugin/plugins/axiom/skills/axiom-security/SKILL.md:216
F-01 Prompt injectionCWE-94 · CWE-1427Layer L2
→ Read: `skills/keychain-diag.md`
Why
asks the agent to read credentials
Fix
—
HIGHPrompt injectionprompt.credential_read · .claude-plugin/plugins/axiom/skills/axiom-security/SKILL.md:219
F-01 Prompt injectionCWE-94 · CWE-1427Layer L2
→ Read: `skills/keychain-ref.md`
Why
asks the agent to read credentials
Fix
—
HIGHPrompt injectionprompt.hide_from_user · .claude-plugin/plugins/axiom/skills/axiom-apple-docs/SKILL.md:209
F-01 Prompt injectionCWE-94 · CWE-1427Layer L2
Do not silently fail — tell the user when Xcode docs aren't available locally and which fallback you used.
Why
asks the agent to act without the user's knowledge
Fix
—
HIGHPrompt injectionprompt.hide_from_user · axiom-codex/skills/axiom-apple-docs/SKILL.md:209
F-01 Prompt injectionCWE-94 · CWE-1427Layer L2
Do not silently fail — tell the user when Xcode docs aren't available locally and which fallback you used.
Why
asks the agent to act without the user's knowledge
Fix
—
HIGHPrompt injectionprompt.hide_from_user · axiom-cursor/skills/axiom-apple-docs/SKILL.md:209
F-01 Prompt injectionCWE-94 · CWE-1427Layer L2
Do not silently fail — tell the user when Xcode docs aren't available locally and which fallback you used.
Why
asks the agent to act without the user's knowledge
Fix
—
HIGHPrompt injectionprompt.override · .claude-plugin/plugins/axiom/agents/iap-auditor.md:120
F-01 Prompt injectionCWE-94 · CWE-1427Layer L2
**Issue**: Fraudulent receipts granted entitlements; jailbreak exploit surface
Why
asks the agent to drop prior instructions or safety
Fix
remove the instruction
HIGHPrompt injectionprompt.override · .claude-plugin/plugins/axiom/agents/iap-auditor.md:237
F-01 Prompt injectionCWE-94 · CWE-1427Layer L2
| Is server-side receipt validation in place for high-value entitlements, or is validation purely client-side? | Weak entitlement enforcement | Jailbreak/emulator bypass grants paid features for free
Why
asks the agent to drop prior instructions or safety
Fix
remove the instruction
HIGHPrompt injectionprompt.override · .claude-plugin/plugins/axiom/agents/security-privacy-scanner.md:118
F-01 Prompt injectionCWE-94 · CWE-1427Layer L2
**Issue**: UserDefaults is unencrypted — accessible via backup extraction and jailbreak
Why
asks the agent to drop prior instructions or safety
Fix
remove the instruction
MEDIUMInventory / provenanceinv.binary · .claude-plugin/plugins/axiom/tools/axbuild
F-13 Inventory / provenanceCWE-1104Layer L0
axbuild
Why
a compiled or binary member cannot be reviewed from source
Fix
ship source, or explain the binary in the README
MEDIUMInventory / provenanceinv.binary · axiom-codex/tools/axbuild
F-13 Inventory / provenanceCWE-1104Layer L0
axbuild
Why
a compiled or binary member cannot be reviewed from source
Fix
ship source, or explain the binary in the README
MEDIUMNetwork egressnet.raw_ip · axiom-codex/skills/axiom-scan-security-privacy/SKILL.md:121
F-04 Network egressCWE-200 · CWE-319Layer L1
**Note**: Exclude `http://localhost`, `http://127.0.0.1`, and documentation strings.
Why
No further detail recorded for this rule.
Fix
—
MEDIUMNetwork egressnet.raw_ip · scripts/codex-lifecycle.py:662
F-04 Network egressCWE-200 · CWE-319Layer L1
'model_provider="axiom_fixture"\nmodel="axiom-synthetic-fixture"\n[features]\nhooks=true\nmulti_agent=true\n[model_providers.axiom_fixture]\nname="Authored transport without inference"\nbase_url="http
Why
No further detail recorded for this rule.
Fix
—
MEDIUMObfuscation / stealthobf.homoglyph · tools/xcprof/cmd_compare.go:176
F-07 Obfuscation / stealthCWE-506 · CWE-94Layer L1
b.WriteString("| function | binary | incl % (base→cur) | self Δpp | ~ms Δ | severity |\n|---|---|---|---|---|---|\n")
Why
No further detail recorded for this rule.
Fix
—
MEDIUMHard-coded secretssecret.generic · .claude-plugin/plugins/axiom/hooks/hook_diagnostics_test.py:18
F-03 Hard-coded secretsCWE-798 · CWE-321Layer L1
SECRET = "…"
Why
No further detail recorded for this rule.
Fix
—
MEDIUMPrompt injectionprompt.conditional_escalation · axiom-codex/skills/axiom-health-check/SKILL.md:106
F-01 Prompt injectionCWE-94 · CWE-1427Layer L2
If the user says "skip X" or "exclude X", remove that auditor from the run list. Acknowledge which auditors were excluded and why.
Why
No further detail recorded for this rule.
Fix
—
LOWInventory / provenanceinv.hidden_file · .mise.toml
F-13 Inventory / provenanceCWE-1104Layer L0
.mise.toml
Why
hidden member outside the usual dotfiles
Fix
review its purpose
LOWInventory / provenanceinv.suspicious_name · scripts/cursor/real-payload.test.ts
F-13 Inventory / provenanceCWE-1104Layer L0
real-payload.test.ts
Why
member named after an attack tool
Fix
remove or justify
LOWCode injectioncode.eval_exec · .claude-plugin/plugins/axiom/hooks/user-prompt-submit_test.py:86
F-02 Code injectionCWE-78 · CWE-94 · CWE-95Layer L1
exec(HOOK_CODE, {"__name__": "__main__", "__file__": HOOK})
Why
evaluates text as code
Fix
remove; use a parser or a dispatch table
LOWCode injectioncode.eval_exec · axiom-mcp/src/tools/binary-exec.test.ts:30
F-02 Code injectionCWE-78 · CWE-94 · CWE-95Layer L1
const r = await exec(NODE, ['-e', "process.stdout.write('hello')"]);
Why
evaluates text as code
Fix
remove; use a parser or a dispatch table
LOWCode injectioncode.eval_exec · axiom-mcp/src/tools/binary-exec.test.ts:36
F-02 Code injectionCWE-78 · CWE-94 · CWE-95Layer L1
const r = await exec(NODE, ['-e', "process.stderr.write('boom');process.exit(3)"]);
Why
evaluates text as code
Fix
remove; use a parser or a dispatch table
LOWCode injectioncode.eval_exec · axiom-mcp/src/tools/binary-exec.test.ts:45
F-02 Code injectionCWE-78 · CWE-94 · CWE-95Layer L1
const r = await exec(NODE, [
Why
evaluates text as code
Fix
remove; use a parser or a dispatch table
LOWCode injectioncode.eval_exec · axiom-mcp/src/tools/binary-exec.test.ts:55
F-02 Code injectionCWE-78 · CWE-94 · CWE-95Layer L1
const r = await exec(NODE, ['-e', "process.stdout.write('x'.repeat(100000))"]);
Why
evaluates text as code
Fix
remove; use a parser or a dispatch table
LOWInsecure cryptocrypto.weak_hash · scripts/codex-output.js:61
F-09 Insecure cryptoCWE-327 · CWE-338Layer L1
return createHash('sha1')
Why
No further detail recorded for this rule.
Fix
—

Audit 0.4.1 · source sha 732d12362944 · 13 Nov 2026Report an issue or request a re-scan

Provenance

Who published it, and what the verdict is pinned to

  • Publishercharleswiltgen GitHub userOwner of the repository this skill is read from
  • Source foldercharleswiltgen/axiomThe exact folder we audited
  • LicenceMITAs the repository states it
  • Commit pinned for this verdict732d12362944A new audit runs when the source changes
  • Repository statusNot archivedLast push 4 days ago
  • Sandboxed runNot runSkills are read, not executed; layer L4 is skipped

What it tells the agent

Full source on GitHub ↗

Its description: Use when fixing or auditing ANY accessibility issue — VoiceOver, Dynamic Type, color contrast, touch targets, WCAG compliance, App Store accessibility review.

---
name: axiom-accessibility
description: Use when fixing or auditing ANY accessibility issue — VoiceOver, Dynamic Type, color contrast, touch targets, WCAG compliance, App Store accessibility review.
license: MIT
---

# Accessibility

**You MUST use this skill for ANY accessibility work including VoiceOver, Dynamic Type, color contrast, WCAG compliance, and UX flow auditing.**

<!-- AXIOM_AUDITOR_INLINE_BEGIN — rewritten for Codex by scripts/build-codex.ts; do not hand-edit -->
> **Auditors are skills here.** Where this router says "Launch `some-auditor` agent", invoke the
> matching Codex skill instead — same procedure, no Claude Code agent required.
>
> Available: `axiom-audit-accessibility`, `axiom-audit-concurrency`, `axiom-audit-swiftdata`, `axiom-audit-swiftui-nav`, `axiom-audit-ux-flow`.
>
> The ones that shell out — builds, tests, simulators, crash symbolication — need shell access to run.
<!-- AXIOM_AUDITOR_INLINE_END -->

## Quick Reference

| Symptom / Task | Reference |
|----------------|-----------|
| VoiceOver labels, hints, navigation | See `skills/accessibility-diag.md` |
| Dynamic Type scaling violations | See `skills/accessibility-diag.md` |
| Dynamic Type on tvOS (Large Text, tvOS 27) | See `skills/accessibility-diag.md` |
| Long-form reading apps (continuous reading, Speak Screen, text navigation) | See `skills/accessibility-diag.md` |
| Captions & subtitle styling in video players (generated subtitles, style preview) | See `skills/accessibility-diag.md` |
| Custom control technique choice (adjustable, passthrough, direct touch) | See `skills/accessibility-diag.md` |
| Accessibility Nutrition Labels | See `skills/accessibility-diag.md` |
| Color contrast (WCAG AA/AAA) | See `skills/accessibility-diag.md` |
| Touch target sizes (< 44x44pt) | See `skills/accessibility-diag.md` |
| Keyboard navigation (iPadOS/macOS) | See `skills/accessibility-diag.md` |
| Button Shapes, Large Content Viewer, Full Keyboard Access, Switch Control | See `skills/accessibility-diag.md` |
| VoiceOver order broken after resize / column collapse; table & grid semantics | See `skills/accessibility-diag.md` |
| Reduce Motion support | See `skills/accessibility-diag.md` |
| Assistive Access (cognitive, iOS 17+) | See `skills/accessibility-diag.md` |
| Accessibility Inspector workflows | See `skills/accessibility-diag.md` |
| UX dead ends, dismiss traps | See `skills/ux-flow-audit.md` |
| Buried CTAs, missing empty states | See `skills/ux-flow-audit.md` |
| Missing loading/error states | See `skills/ux-flow-audit.md` |
| Deep link dead ends | See `skills/ux-flow-audit.md` |
| Accessibility dead ends (gesture-only) | See `skills/ux-flow-audit.md` |
| watchOS-specific (VoiceOver rotor on Digital Crown, AssistiveTouch, Double Tap) | See `skills/watchos-a11y.md` |

## Cross-Suite Routes

- Full watchOS development context → See axiom-watchos
- Live accessibility validation on the simulator (set toggles, assert announcements) → `simulator-tester` agent + `xcui` — see axiom-tools (skills/xcui-ref.md)
- Asserting VoiceOver speech in an XCUITest (`XCUIDevice.shared.voiceOverService`, `OS27`) → axiom-testing (skills/ui-testing.md Pattern 7)

## Decision Tree

```dot
digraph accessibility {
    start [label="Accessibility issue" shape=ellipse];
    what [label="What type?" shape=diamond];

    start -> what;
    what -> "skills/accessibility-diag.md" [label="VoiceOver/labels/hints"];
    what -> "skills/accessibility-diag.md" [label="Dynamic Type"];
    what -> "skills/accessibility-diag.md" [label="color contrast"];
    what -> "skills/accessibility-diag.md" [label="touch targets"];
    what -> "skills/accessibility-diag.md" [label="keyboard nav"];
    what -> "skills/accessibility-diag.md" [label="resizable window / desktop-class"];
    what -> "skills/accessibility-diag.md" [label="Reduce Motion"];
    what -> "skills/accessibility-diag.md" [label="Assistive Access"];
    what -> "skills/accessibility-diag.md" [label="custom control techniques"];
    what -> "skills/accessibility-diag.md" [label="continuous reading / text navigation"];
    what -> "skills/accessibility-diag.md" [label="captions / subtitle styling"];
    what -> "skills/accessibility-diag.md" [label="App Store prep / Nutrition Labels"];
    what -> "skills/ux-flow-audit.md" [label="UX dead end/dismiss trap"];
    what -> "skills/ux-flow-audit.md" [label="missing states"];
    what -> "skills/watchos-a11y.md" [label="watchOS VoiceOver / AssistiveTouch / Double Tap"];
    what -> "accessibility-auditor" [label="automated scan" shape=box];
}
```

1. ANY VoiceOver, Dynamic Type (including tvOS Large Text), contrast, touch target, or WCAG issue → `skills/accessibility-diag.md`
2. Assistive Access (cognitive disabilities, iOS 17+) → `skills/accessibility-diag.md`
3. Accessibility Nutrition Labels, or an App Review note about accessibility → `skills/accessibility-diag.md`
4. Reading app: VoiceOver stops at paragraphs/pages, Speak Screen halts → `skills/accessibility-diag.md`
5. UX dead ends, dismiss traps, buried CTAs, missing states → `skills/ux-flow-audit.md`
6. watchOS-specific accessibility (rotor on Digital Crown, AssistiveTouch, Double Tap) → `skills/watchos-a11y.md`
7. Want automated accessibility scan? → `accessibility-auditor` agent or `axiom-audit-accessibility`

## Automated Scanning

**Accessibility audit** → `axiom-audit-accessibility`
- VoiceOver labels and hints
- Dynamic Type violations
- Color contrast failures
- WCAG compliance scanning

**UX flow audit** → `axiom-audit-ux-flow`
- Dead-end views, dismiss traps
- Buried CTAs, missing empty/loading/error states
- Deep link dead ends, accessibility dead ends

## Critical Patterns

#### Image Accessibility

- Use `Image(decorative: "photo")` for purely decorative images — automatically hidden from VoiceOver (equivalent to `accessibilityHidden(true)` but semantically clearer)
- Use `accessibilityInputLabels()` for buttons with complex or changing labels — improves Voice Control acc

Excerpt, first 6,000 characters of SKILL.md, verbatim from the audited commit. It is text from someone else’s repository, quoted so you can judge it before your agent reads it.

Install

Pinned to the audited commit 732d123

git clone https://github.com/Example/PackageB
pip install -r requirements.txt
git clone https://github.com/yourname/package.git
claude mcp add --transport stdio --scope user asc-mcp -- ~/.mint/bin/asc-mcp
claude mcp add --transport stdio xcode -- xcrun mcpbridge

Commands as the repository documents them: shown, never run by us. Our verdict covers commit 732d12362944 only.

Works in which agents?

What its own text says, not a test

AgentStatusMore
Claude CodeMentioned in its text, not testedSkills for Claude Code · About Claude Code
Claude DesktopMentioned in its text, not testedSkills for Claude Desktop · About Claude Desktop
CodexMentioned in its text, not testedSkills for Codex · About Codex
GitHub CopilotMentioned in its text, not testedSkills for GitHub Copilot · About GitHub Copilot
CursorMentioned in its text, not testedSkills for Cursor · About Cursor
Gemini CLIMentioned in its text, not testedSkills for Gemini CLI · About Gemini CLI
“Mentioned” means the skill’s text names the agent. We have not run it there.

Audit history

Every audit, every source change

  1. ✕BLOCK F 57 / 100

    first audit · source 732d12362944

README badge

For maintainers · shows the latest audit

ModelsAtlas: BLOCK FMODELSATLASBLOCK · F
[![ModelsAtlas: BLOCK F](https://modelsatlas.com/skills/charleswiltgen/axiom-swiftdata/badge.svg)](https://modelsatlas.com/skills/charleswiltgen/axiom-swiftdata/)

The badge shows our verdict and grade for this skill, and links back to this page. It changes when the audit does.

Similar skills

Add this badge to your README

[![ModelsAtlas: BLOCK F](https://modelsatlas.com/skills/charleswiltgen/axiom-swiftdata/badge.svg)](https://modelsatlas.com/skills/charleswiltgen/axiom-swiftdata/)

Questions about axiom-swiftdata

What does the axiom-swiftdata skill do?
Battle-tested Claude Code skills for modern xOS (iOS, iPadOS, watchOS, tvOS) development
Is axiom-swiftdata safe to install?
No — not without reading the findings first. The audit graded it F (57/100) and found 10 critical or high issues in the source. Each one is listed on this page with the file and line it is on.
What can axiom-swiftdata access on my machine?
The audit observed that it reaches the network and runs shell commands. Each of those is consistent with what it says it does. Secrets in the source: found — see the findings.
Which assistants does axiom-swiftdata work with?
Its documentation mentions claude-code, claude-desktop, codex, copilot, cursor and gemini-cli. That is what the text claims, not a compatibility test we ran.
How current is this page?
The grade is for one exact copy of the source (732d12362944), read on 2026-11-13. The repository is watched, and a new audit runs when it changes — this is the first audit.

Sources

Spotted something wrong? Report a correction · Methodology

Keep going in your AI

Opens with this page as the source. We add nothing else to the prompt.