DE

detect-silent-processing-failures

Identifies suspicious pipeline outputs even when no exceptions occur by validating metadata and file structures.

Install

mkdir -p .claude/skills/detect-silent-processing-failures && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/18496" && unzip -o skill.zip -d .claude/skills/detect-silent-processing-failures && rm skill.zip

Installs to .claude/skills/detect-silent-processing-failures

Activation

This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.

Use to detect suspicious processor outputs even when no exception is thrown. Keywords: silent failure, metadata validation, glb checks, anomaly detection.
154 chars✓ has a “when” trigger
Intermediate

Key capabilities

  • Report metadata accounting inconsistencies
  • Validate referenced models for existence and size
  • Flag suspicious mesh and bounds values
  • Run checks in strict mode for more rigorous validation
  • Provide a compact summary of findings
  • Offer detailed per-item evidence for flagged issues

How it works

This skill executes a Python script to check processed outputs for silent failures by validating metadata, model references, and mesh/bounds values.

Inputs & outputs

You give it
Metadata JSON file from processor outputs
You get back
Reports on metadata inconsistencies, missing model files, and suspicious mesh/bounds values

When to use detect-silent-processing-failures

  • Validating pipeline outputs
  • Checking metadata consistency
  • Flagging suspicious mesh or bounds data

About this skill

Detect Silent Processing Failures Skill

Use this skill after processor runs to identify suspicious outcomes.

Outcomes

  • Metadata accounting inconsistencies are reported.
  • Referenced models are validated for existence and size.
  • Suspicious mesh and bounds values are flagged.

Core Command

python3 scripts/check_processed_outputs.py --metadata data/outputs/metadata.json

Strict mode:

python3 scripts/check_processed_outputs.py --metadata data/outputs/metadata.json --strict

Warning Heuristics

  • hhh_count differs from hhh_inspected.
  • model_count + export_failed_count differs from hhh_inspected.
  • Item references missing or empty model file.
  • Invalid mesh_bounds structure.

Agent Guidance

  • Default to warnings only unless strict mode requested.
  • Provide a compact summary and detailed per-item evidence.
  • Keep checks deterministic and schema-version aware.

When not to use it

  • When an exception is already thrown by the processor
  • When only basic error detection is needed
  • When schema-version awareness is not a concern

Limitations

  • Relies on specific warning heuristics for detection
  • Default behavior is warnings only unless strict mode is requested
  • Checks are deterministic and schema-version aware

How it compares

This skill proactively identifies subtle processing failures that do not throw exceptions, providing a deeper level of data integrity validation than relying solely on error codes.

Compared to similar skills

detect-silent-processing-failures side by side with the closest alternatives in the catalog.

SkillInstallsUpdatedSafetyDifficulty
detect-silent-processing-failures (this skill)04moReviewIntermediate
python-testing-patterns772moReviewIntermediate
backtesting-frameworks172moNo flagsAdvanced
python-repl64moReviewBeginner

Try saying

Example prompts that trigger this skill in your AI assistant.

Search skills

Search the agent skills registry