# Jobs To Be Done

<!-- BEGIN:jobs — Do not edit. Generated from products/*/package.json. -->

<job user="Engineering Leaders" goal="Define the Engineering Standard">

## Engineering Leaders: Define the Engineering Standard

**Trigger:** Two managers disagree on what 'senior' means — neither can point to
a written definition.

**Big Hire:** Help me define what good engineering means so roles have clear,
defensible expectations. → **Map**

**Little Hire:** Help me define or update a role's expectations knowing
structural mistakes get caught. → **Map**

**Competes With:** tribal knowledge; borrowed frameworks; per-manager intuition;
tolerating the ambiguity.

**Forces:**
- **Push:** Inconsistent expectations produce contested promotions.
- **Pull:** A shared definition that makes quality visible and discussable.
- **Habit:** Each manager carries a private mental model of what 'good' means.
- **Anxiety:** Getting definitions wrong feels worse than having none.

**Fired When:** definitions drift from practice; or a reorg removes the mandate.

</job>

<job user="Engineering Leaders" goal="Measure Engineering Outcomes">

## Engineering Leaders: Measure Engineering Outcomes

**Trigger:** Quarterly review is due and the only data is ticket counts —
singling out individuals, not illuminating the system.

**Big Hire:** Help me demonstrate engineering progress without making
individuals feel surveilled. → **Landmark**

**Little Hire:** Help me tell whether culture investments are working before the
next budget cycle. → **Landmark**

**Competes With:** sprint velocity; ticket counts; annual reviews; asking
managers 'how's the team doing?'; not measuring.

**Forces:**
- **Push:** Pressure to show ROI with no way to separate signal from noise.
- **Pull:** System-level trends that show direction without naming individuals.
- **Habit:** Equating velocity and throughput with team health.
- **Anxiety:** Measurement feels like surveillance regardless of intent.

**Fired When:** metrics get used punitively; or leadership turnover
deprioritizes measurement.

</job>

<job user="Engineering Leaders" goal="Staff Teams to Succeed">

## Engineering Leaders: Staff Teams to Succeed

**Trigger:** A post-mortem surfaces the same skill gap that caused the last
incident — nobody saw it before staffing.

**Big Hire:** Help me make staffing decisions I can defend by seeing what each
role requires. → **Summit**

**Little Hire:** Help me spot capability gaps and check whether a candidate
fills them. → **Summit**

**Competes With:** résumé screening; gut feel; copying a team that seems to
work; accepting gaps as inevitable.

**Forces:**
- **Push:** Capability gaps appear as incidents, never in advance.
- **Pull:** Confidence that a staffing change strengthens the team as a system.
- **Habit:** Hiring on résumés and gut rather than team composition analysis.
- **Anxiety:** Modeling human capability feels reductive.

**Fired When:** the skill taxonomy drifts from actual roles; or a hiring freeze
removes staffing pressure.

</job>

<job user="Empowered Engineers" goal="Be Prepared and Productive">

## Empowered Engineers: Be Prepared and Productive

**Trigger:** Walking into a meeting cold because context was scattered across
email, Slack, and last week's notes.

**Big Hire:** Help me keep track of people, projects, and threads without
depending on memory. → **Outpost**

**Little Hire:** Help me walk into every meeting already oriented. → **Outpost**

**Competes With:** morning email scan; re-reading calendar invites; personal
notes; asking colleagues for context.

**Forces:**
- **Push:** Context-switching compounds; starting cold wastes everyone's time.
- **Pull:** Continuous awareness without continuous effort — context assembled.
- **Habit:** Morning routines and tracking systems that feel 'good enough.'
- **Anxiety:** Delegating awareness to a system feels like losing control.

**Fired When:** the briefing becomes noise; or simpler tools absorb the
preparation.

</job>

<job user="Empowered Engineers" goal="Get Judgment Grounded in the Standard">

## Empowered Engineers: Get Judgment Grounded in the Standard

**Trigger:** The promotion conversation ended with 'not yet' but no specifics,
and agent output gets reviewed line by line because nobody can tell whether it
meets the bar.

**Big Hire:** Help me get career guidance and output review grounded in my
organization's standard, not impressions or generic advice. → **Guide**

**Little Hire:** Help me ask a growth question or get a second opinion on a
deliverable before approving it. → **Guide**

**Competes With:** annual reviews; mentor conversations; reviewing every line;
trusting uncritically; hoping the next project makes readiness obvious.

**Forces:**
- **Push:** Career conversations and agent review both run on subjective
  judgment with no shared evidence base.
- **Pull:** A second opinion grounded in the standard, whether on a career or a
  deliverable.
- **Habit:** Waiting for annual feedback and reviewing every line by hand.
- **Anxiety:** Structured assessment might confirm being further behind than
  assumed.

**Fired When:** assessments disconnect from promotion and approval decisions; or
a reorg makes the standard feel irrelevant.

</job>

<job user="Empowered Engineers" goal="See What's Expected of Humans and Agents">

## Empowered Engineers: See What's Expected of Humans and Agents

**Trigger:** 'Meets expectations' on the review form has no definition anyone
can point to, and an agent's work was just rejected for following generic
practices instead of the organization's.

**Big Hire:** Help me see what's expected of me at my level and of the agents I
configure, from one shared standard. → **Pathway**

**Little Hire:** Help me generate a job definition, a level comparison, or an
agent profile from discipline, track, and level. → **Pathway**

**Competes With:** asking peers; inferring from manager comments; custom system
prompts; copying another team's config; living with ambiguity.

**Forces:**
- **Push:** Ambiguous expectations create review anxiety for humans and rejected
  work from agents.
- **Pull:** One standard that produces role definitions for people and profiles
  for agents.
- **Habit:** Inferring expectations from peer behavior and hand-tuning agent
  prompts per team.
- **Anxiety:** Looking up the standard might confirm gaps easier left ignored.

**Fired When:** defined expectations diverge from actual promotions; or
configuration overhead exceeds the quality gain.

</job>

<job user="Platform Builders" goal="Build Agent-Capable Systems">

## Platform Builders: Build Agent-Capable Systems

**Trigger:** Building an agent that needs structured knowledge or typed
contracts, with no way to prove the next change improves outcomes.

**Big Hire:** Help me give humans and agents shared capabilities through the
same interface, with tooling to prove changes improved outcomes. → **Gear**

**Little Hire:** Help me query graphs, run vector search, expose services as MCP
tools, or chart agent metrics without building infrastructure. → **Gear**

**Competes With:** hand-rolled retrieval; bespoke wrappers; custom eval
harnesses; monolithic platforms; shipping without measuring.

**Forces:**
- **Push:** Every agent project rebuilds plumbing, and improvements are claimed
  but never measured.
- **Pull:** Interoperable libraries and reproducible evals that work standalone
  or together.
- **Habit:** Copy-pasting infrastructure and assuming changes helped because the
  prompt looks better.
- **Anxiety:** Shared libraries mean losing control over the interface contract.

**Fired When:** the libraries diverge from how agents consume them; or eval
overhead exceeds the value of knowing.

</job>

<job user="Teams Using Agents" goal="Run a Continuously Improving Agent Team">

## Teams Using Agents: Run a Continuously Improving Agent Team

**Trigger:** Agents are shipping work but nobody can tell whether the team is
getting better — the only feedback loop is reading every diff.

**Big Hire:** Help me run an autonomous, continuously improving development team
that plans, ships, studies its own traces, and acts on findings. → **Kata**

**Little Hire:** Help me onboard a Kata installation that runs the
Plan-Do-Study-Act loop without per-team prompt engineering. → **Kata**

**Competes With:** bespoke per-agent system prompts; manual orchestration
scripts; not measuring agent outcomes; abandoning agent investment after a
failed pilot.

**Forces:**
- **Push:** Agent regressions are silent until users complain.
- **Pull:** A closed loop that surfaces what improved and what regressed,
  grounded in evidence.
- **Habit:** Treating each agent run as a one-off rather than an iteration.
- **Anxiety:** Autonomy might amplify bad patterns faster than humans can
  intervene.

**Fired When:** the autonomous loop becomes harder to operate than direct
prompting; or organizational policy bans autonomous agent execution.

</job>

<!-- END:jobs -->