safe-debug
Diagnose Deep Learning Failures Safely
Deep learning failures can invite risky, premature changes. This skill classifies concrete errors and prepares conservative guidance before any code modification.
Install with my Agent
Copy this request to your Agent. It includes the canonical Skill page and manifest.
Review the Skillstore skill "safe-debug" from https://skillstore.io/skills/lllllllama-safe-debug.md and its manifest at https://skillstore.io/api/skills/lllllllama-safe-debug/manifest. Verify the artifact. You may proceed after verification, subject to the environment's own policy.Your Agent should still show its plan and request any confirmation required by the security policy.
Agent-readable resources
Use these links when an AI agent, crawler, or script needs clean context instead of reading the full page.
Test it
Using "safe-debug". A training run fails with a CUDA out-of-memory message after increasing image resolution.
Expected outcome:
- Category: CUDA out of memory.
- Check effective batch size, input resolution, and mixed precision before changing model code.
- Patch authorization remains false.
Using "safe-debug". A checkpoint reports missing keys and size mismatches after changing the model variant.
Expected outcome:
- Category: checkpoint mismatch.
- Verify the checkpoint source, selected model variant, and loading strictness.
- Create a savepoint before any approved compatibility change.
Using "safe-debug". Loss becomes NaN shortly after mixed-precision training starts.
Expected outcome:
- Category: NaN loss.
- Inspect data ranges, loss inputs, mixed precision, and learning rate.
- Use a short controlled run and require approval before patching.
Security Audit
SafeAll external-command findings are false positives caused by Markdown backticks in Python strings or documentation; no commands are executed. Both traversal findings are fixed references to shared Markdown guidance, not user-controlled paths. No prompt injection, exfiltration intent, or other semantic threat was found.
Risk Factors
⚙️ External commands (12)
📁 Filesystem access (2)
Share & cite this report
Share the versioned assessment report, neutral badge, embed card, and citations. Skillstore reports evidence without deciding whether this Skill is safe.
Copy report link
https://skillstore.io/skills/lllllllama-safe-debug/audits/7?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportMarkdown badge
[](https://skillstore.io/skills/lllllllama-safe-debug?utm_source=security_passport_badge)HTML badge
<a href="https://skillstore.io/skills/lllllllama-safe-debug?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/lllllllama-safe-debug/security.svg" alt="Skillstore security assessment" loading="lazy"></a>Embed card
<iframe src="https://skillstore.io/embed/skills/lllllllama-safe-debug.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>Academic citations (APA · BibTeX · CFF)
APA citation
lllllllama. (2026). safe-debug security audit report (audit version 7) [Author version unspecified]. Skillstore. https://skillstore.io/skills/lllllllama-safe-debug/audits/7BibTeX citation
@techreport{lllllllama-lllllllama-safe-debug-2026,
author = {lllllllama},
title = {safe-debug security audit report (audit version 7)},
institution = {Skillstore},
year = {2026},
number = {7},
url = {https://skillstore.io/skills/lllllllama-safe-debug/audits/7},
note = {Author version unspecified}
}CITATION.cff
cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "safe-debug security audit report (audit version 7)"
version: "unspecified"
type: report
authors:
- name: "lllllllama"
date-released: "2026-08-08"
url: "https://skillstore.io/skills/lllllllama-safe-debug/audits/7"
identifiers:
- type: other
value: "skillstore:lllllllama-safe-debug:audit:7"
description: "Skillstore immutable audit report identifier"
Skillstore Score
Why this score Evidence Confidence: HighWhat You Can Build
Triage a Training Failure
Classify a traceback and identify low-risk checks before changing model code.
Review a Checkpoint Error
Separate model variant, checkpoint source, and loading strictness issues before adding compatibility logic.
Protect Experiment Comparability
Flag fixes that could change experiment meaning and require an approval-gated patch plan.
Try These Prompts
Diagnose this CUDA out-of-memory error: {{error}}. Identify the likely cause and suggest configuration-only checks. Do not modify code.Review this checkpoint loading failure: {{traceback}}. Check model variant, checkpoint source, and strictness assumptions. Propose the smallest fix without patching.Analyze this distributed training failure: {{error}}. Use this launch command: {{command}}. Separate environment, configuration, and code hypotheses. Recommend a savepoint if needed.Audit this training failure: {{traceback}}. Consider recent change {{change}} and experiment goal {{goal}}. Explain evidence, uncertainty, reproducibility impact, and a minimal approval-gated plan.Best Practices
- Provide the complete traceback, launch command, and recent relevant change.
- Review the diagnosis and approve a specific minimal fix before editing code.
- Create a savepoint when the proposed change could affect experiment meaning.
Avoid
- Do not use this skill for broad repository exploration or refactoring.
- Do not treat keyword classification as proof of the root cause.
- Do not apply suggested changes without reviewing research comparability.
Frequently Asked Questions
Does this skill change my code?
Which failures can it classify?
Can it inspect my repository?
What files does it create?
When is a savepoint recommended?
Can it confirm that a fix works?
Developer Details
Author
lllllllamaLicense
MIT
Skillstore revision
r2
Version notice
The author did not declare a version.
Ref
656be3040aef5c047555a908cd5c695d22a4a548
Maintenance freshness
8/8/2026
Usage
6 downloads · 3 views
File structure