| name | kata-interview |
| description | Conduct a JTBD switching interview to test a Forward Impact product. Pick one of the product's Jobs To Be Done, build the persona from that JTBD entry alone, hand the job to the agent at the public website, and capture findings as GitHub issues classified against the chosen job.
|
Switching Interview
You are the supervisor in a fit-eval supervise relay running a JTBD
switching interview: an agent, briefed only with a persona derived from a
chosen Job To Be Done, tries to get that job done using a Forward Impact
product they encounter cold at the public website. The agent is in an
isolated workspace with no monorepo access. You run in the monorepo root
with full access to JTBD.md, the synthetic data/ from fit-terrain build, the supabase CLI, and project context โ use that to stage the
workspace, craft the persona, and verify findings, but never leak.
When to Use
- You are supervising the
kata-interview workflow via fit-eval supervise.
- The task may include
Product: and/or Job: overrides; otherwise pick.
This skill is supervisor-initiated, not part of scheduled runs.
LLM Availability
ANTHROPIC_API_KEY is present in the shell โ libconfig reads it.
LLM-backed products (Guide, Outpost) should work without the agent
configuring an API key. If the agent is asked to supply a key, that is a
bug โ the zero-config promise is broken. Do not tell the agent the
key is pre-configured.
Checklists
<read_do_checklist goal="Protect the interview before briefing the agent">
</read_do_checklist>
<do_confirm_checklist goal="Close the interview cleanly">
</do_confirm_checklist>
Process
Step 0: Read Memory
Per the agent profile: own summary, current week's log, teammates'
summaries. Bias product selection toward products not interviewed recently.
Step 1: Pick the Product
If the task includes Product:, use it. Otherwise pick one of the products
under products/ that has a <job> entry in JTBD.md.
Step 2: Pick the Job
Read JTBD.md. Find every <job> entry whose Big Hire or Little
Hire line names the chosen product (e.g. โ **Guide, Landmark**). If the
task includes Job:, match it against the goal= attribute; otherwise
pick one. Record the full block: user, goal, Trigger, Big Hire, Little
Hire, Competes With, Forces (Push, Pull, Habit, Anxiety), Fired When.
Step 3: Stage the Agent Workspace
The workflow has run bunx fit-terrain build and installed supabase
globally. Copy the subset the chosen product needs into $AGENT_CWD (adjust
for your installation's products):
| Product | Stage into $AGENT_CWD |
|---|
| Guide, Outpost | nothing |
| Pathway | data/pathway/ |
| Map, Landmark | data/pathway/ and data/activity/ |
| Summit | data/pathway/ and data/activity/raw/activity/summit.yaml (as summit.yaml at root) |
Use cp -r data/pathway "$AGENT_CWD/data/pathway" and similar.
Step 4: Craft the Persona
Write $AGENT_CWD/CLAUDE.md using only fields from the chosen JTBD
entry. Do not name the product โ the persona arrives at the website without
foreknowledge. Template:
You are a <user> with a job to be done: <goal>.
## Trigger
<Trigger>
## Forces
- **Push:** <push>
- **Pull:** <pull>
- **Habit:** <habit>
- **Anxiety:** <anxiety>
## What you currently use
<Competes With list โ your current alternatives>
## Hire / Fire
You'll hire something that helps you: <bigHire then littleHire>.
You'll abandon it when: <firedWhen>.
## How to act
Try to get this job done. Start at https://www.forwardimpact.team โ the
only entry point. Follow docs as written; don't seek workarounds. Install
from npm as a normal user would; don't clone any monorepo. Note friction
in your final output โ do not write findings to files.
Step 5: Initiate the Session
Your first response is the agent's initial message. Short and in character:
Welcome. Read CLAUDE.md for who you are and what you need to get done.
Start at https://www.forwardimpact.team. Get the job done. Report
findings in your final output, not in files.
If the task carries steering not matching Product: / Job:, append it.
Step 6: Supervise
| Agent State | Your Response |
|---|
| Making progress | Short encouragement |
| Stuck on a specific step | Answer the specific question, in character |
| Going down a dead end | Nudge toward the documented path |
| Looping without progress | Targeted guidance |
| Job done or abandoned | Proceed to Step 7 |
Use monorepo access to verify observations โ but do not feed verification
back to the agent.
Step 7: Signal Completion
When the persona has gotten the job done or clearly abandoned it, write
INTERVIEW_COMPLETE and continue with post-interview work in the same
turn. Do not stop.
Step 8: Capture Findings
Review the agent's output. For each distinct finding, note against the
JTBD: was the Big Hire reached? Little Hire experienced? Did
Anxiety land? Did Competes With look more attractive? Did any
Fired When condition surface?
Classify each for action:
| Category | Criteria | Action |
|---|
| Bug | Crashes, errors, wrong output | Create bug issue |
| Product-aligned | Missing feature serving the vision | Create feature issue |
| Documentation | Unclear, missing, or outdated docs | Create docs issue |
| Out of scope | Not actionable or outside the product | Skip โ note in report |
For each actionable finding: extract; search for duplicates; create a new
issue or comment on a matching one (templates in
../kata-product-issue/references/templates.md ยง New Issues from User
Testing) naming the JTBD job (<user>: <goal>) in the body; add the
finding to the report table with its issue number.
Step 9: Report
Final summary: product and job; whether the persona got it done; which JTBD
forces materialised; table of findings and issues created or updated.
Memory: what to record
Append to the current week's log:
- Product โ interviewed
- Job โ
<user>: <goal>
- Outcome โ done / abandoned / partial
- Forces observed โ Push/Pull/Habit/Anxiety/Competes/Fired
- Issues created or updated โ numbers and categories
- Metrics โ Append one row per run to
wiki/metrics/{skill}/ per
references/metrics.md. See KATA.md ยง Metrics for recording eligibility.
