Claude-skill-registry load-anomaly-guard

Detects unsafe training load spikes (>20-30% week-over-week) and emits safety flags. Use in nightly background jobs or when reviewing weekly training volume with conservative adjustment recommendations.

install

source · Clone the upstream repo

git clone https://github.com/majiayu000/claude-skill-registry

Claude Code · Install into ~/.claude/skills/

T=$(mktemp -d) && git clone --depth=1 https://github.com/majiayu000/claude-skill-registry "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/data/load-anomaly-guard" ~/.claude/skills/majiayu000-claude-skill-registry-load-anomaly-guard && rm -rf "$T"

manifest: skills/data/load-anomaly-guard/SKILL.md

When Claude should use this skill

Nightly background check on training data
Immediately after a high-intensity or long run is logged
When analyzing weekly training load patterns for safety issues

Invocation guidance

Provide recent
```
TrainingHistory
```
, planned
```
Plan
```
window, and any injury flags.
Compute week-over-week changes and monotony; flag spikes > deterministic caps.
Suggest adjustments (rest/swaps) and emit
```
SafetyFlag[]
```
.

Input schema

See

references/input-schema.json

Output schema

See

references/output-schema.json

Integration points

Background job: nightly cron.
API:
```
v0/app/api/plan/load-guard
```
(new) returning flags + suggested adjustments.
UI: Badge on Plan/Today screens; push/email via
```
v0/lib/email.ts
```
.

Safety & guardrails

If spike >20–30% week-over-week, emit
```
load_spike
```
and recommend rest or reduced volume.
If injury signals present, bias toward
```
rest-day
```
adjustments.
No medical diagnosis; advise professional consult on repeated spikes or pain.

Telemetry

Emit

ai_skill_invoked

ai_safety_flag_raised

, and optionally

ai_adjustment_applied

when suggestions are auto-applied.