← ClaudeAtlas

weekly-coaching-patternslisted

Interval scheduled check — detects multi-session cardiac-drift trends across recent compiled activity notes. Runs weekly via the scheduled-checks routine; findings are routed through the proposal pipeline as Evidence Source: scheduled-check/weekly-coaching-patterns.
gtapps/claude-code-hermit · ★ 59 · AI & Automation · score 81
Install: claude install-skill gtapps/claude-code-hermit
# Weekly Coaching Patterns Scheduled-check skill: reads the last 4 `compiled/activity-*.md` artifacts for steady sessions and detects whether cardiac drift is trending upward over time. Returns a fixed findings block for `reflect-scheduled-checks` to classify and route. **Contract:** idempotent, read-only, no self-scheduling, short-running. Returns findings or silence — never creates proposals itself. ## Steps 1. **Glob activity notes.** Use `Glob` on `.claude-code-hermit/compiled/activity-*.md`. If no files match, output the zero-findings block (Step 5, no-trend path) and stop. 2. **Read and filter.** For each matched file, Read it. Keep only entries where the YAML frontmatter has `type: activity-note` AND `session_kind: steady`. Sort by `created` date descending (most recent first). Take the 4 most recent steady sessions. If fewer than 4 steady sessions exist across all artifacts, output the zero-findings block and stop. This is expected on weeks with predominantly interval or strength training — insufficient data is not an error. 3. **Extract cardiac-drift values.** Working oldest-to-newest (reverse the most-recent-first list), for each of the 4 steady sessions parse the body for a line starting with `Cardiac drift:`. Extract the **signed** integer bpm value, preserving the sign (`+12` → `+12`, `-5` → `-5`). A negative value means HR fell over the session (an improving signal) and must compare as less than any positive value — never strip the sign to magnitude. T