Generated skill
Skill dstevenson1981/jtbd-agent-builder/engine/agent-factory/templates/generated-skill
Hire an agent, not a prompt. Give it a business outcome; it diagnoses your system from real data, finds the constraint, and builds self-improving agents with evals and graduated autonomy. Runs in Claude Code and Codex.
npx -y skills add dstevenson1981/jtbd-agent-builder --skill generated-skillAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
<Written for triggering: what job this performs, for which outcome, touching which systems, and the phrasings a user would actually say. No marketing copy.>
SKILL.md
1.3 KB, 287 tokens by cl100k_base, as published. Nobody here has run it
<Job name>
<One paragraph: the job, the metric contract in one line, and where the full contract lives.>
Contract: this skill exists to move <metric> by <Δ> within <window>. Contract, evidence and
current autonomy rungs: JOB.md, evals/, ladder.json (same directory). Check ladder.json
before every action; never write above the capability's current rung — refuse and propose
promotion instead.
Workflow
- <Imperative step — gather inputs from the declared systems>
- <Imperative step — do the work, referencing evidence/voice/format standards>
- Self-check against the rubric in
evals/before producing anything. - For each external action: confirm rung in
ladder.json; at L1 emit a draft, at L2 request approval, at L0 log to shadow output only. - Record every human correction verbatim in
proposals/inbox.md— it becomes an eval.
Escalation
Escalate to a human (do not guess) when: <ambiguity conditions from rejection cases>.
Improvement
This skill does not edit itself. Misses and corrections follow IMPROVEMENT.md: failing eval
first, smallest diff to proposals/, human merges.
What ships with it
Read from the repository
Just SKILL.md. No reference files, no scripts.