autonomous-loop
Iterate until success or limit, composing existing skills with Jidoka integration
Generate breadcrumb headers/footers with takt time tracking and muda metrics
$ npx -y skills add WellApp-ai/Well --skill session-status --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/session-statusContext preview
The summary Claude sees to decide when to auto-load this skill.
Generate breadcrumb headers/footers with takt time tracking and muda metrics
name: session-status description: Generate breadcrumb headers/footers with takt time tracking and muda metrics
Generate consistent navigation breadcrumbs showing current position, takt time, and session metrics.
Collect current session state:
| Data | Source | |------|--------| | Feature name | From init context or branch name | | Current phase | DIVERGE / CONVERGE / DEFINE / AGENT | | Loop count | Incremented on each user feedback cycle | | Progress | Validated items / total items | | Duration | Time since phase started | | Status | G (flow) / Y (waiting) / R (stopped) |
Compare duration against targets from `03-shared.mdc`:
| Phase | Target | Warning | Stop | |-------|--------|---------|------| | DIVERGE (per loop) | 15min | 30min | 60min | | CONVERGE | 20min | 40min | 90min | | DEFINE | 10min | 20min | 45min | | Commit (each) | 10min | 20min | 30min |
Determine suffix:
Generate single-line breadcrumb:
**Format:** `[STATUS] | [Feature] | [PHASE_PROGRESS] | L[N] | [X]/[Y] | [TIME]`
**Examples by phase:**
**Progress bar:** 10 chars, `#` filled, `.` empty, proportional to completion.
Generate context-specific footer:
--- Next: [action prompt]
**Dynamic prompts by context:**
| Context | Footer Prompt | |---------|---------------| | DIVERGE with DIG | `Provide OK/KO/DIG for: #4, #6, #7` | | Gate 1 passed | `All validated. Say "converge" to proceed.` | | Gate 2 | `Choose option: A / B / C` | | Gate 3 | `Approve phasing: OK / REORDER / SPLIT` | | AGENT | `Commit 2/5 in progress...` | | Jidoka stop | `Reply with A, B, C, or D` | | Autonomous | `Iteration 3/5...` |
Update session metrics in memory:
| Metric | Tracking Rule | |--------|---------------| | Loop count | Increment on user feedback | | Waiting time | Time between user responses | | Rework | Items DIG'd 3+ times | | RED count | qa-commit failures | | Escalations | Jidoka Tier 2/3 events |
## Session Status **Header:** [formatted header string] **Footer:** [formatted footer string] ### Metrics (Silent) - Duration: [N]min - Loops: [N] - Rework: [N] - Waiting: [N]min - RED count: [N]
**Header:**
G | Collaboration | DIVERGE ###....... | L3 | 5/9 | 18m
**Footer:**
--- Next: Provide OK/KO/DIG for: #4, #6, #7
**Header:**
G | Collaboration | DIVERGE OK | CONVERGE #......... | L1 | 22m
**Footer:**
--- Next: All wireframes validated. Say "converge" to proceed.
**Header:**
G | Collaboration | AGENT | C2/5 | 45m
**Footer:**
--- Next: Implementing Commit 2: Add workspace membership entity...
**Header:**
R | Collaboration | AGENT | JIDOKA STOP
**Footer:**
--- Next: Reply with A (different approach), B (skip), C (pause), or D (abort)
This skill is invoked by:
Auto-invoked by modes. Manual trigger: "show session status"
| Tool | Purpose | |------|---------| | (none) | Pure calculation, no external tools |
Iterate until success or limit, composing existing skills with Jidoka integration
Generate and maintain BPMN 2.0 diagrams linked to Gherkin scenarios
Research best-in-class products using Browser MCP and WebSearch
Systematic debugging with MCP integration, auto-invoke from qa-commit, Phase 7 Harden