Use it when checking whether what was agreed actually happened.
2 rolesRead moreYour agents already know the job.
A skill is a working playbook — how to review a pipeline deal by deal, close the month so the numbers hold, read an ad account without double-counting. Each one was written around a specific way the work goes wrong, and the agent loads it when the task calls for it. Every one we ship is below.
SHOWING 1–9 OF 9
Use it when deciding whether to do the work or hand it to a teammate.
2 rolesRead moreUse it when opening a role, or asking whether to hire at all.
2 rolesRead moreUse it when asked how the team is doing, or why throughput dropped.
2 rolesRead moreUse it when a role needs an interview process designed.
1 roleRead moreUse it when planning someone's first 30 days, human or agent.
1 roleRead moreUse it when a role needs a pay range or an offer number.
1 roleRead moreUse it when a CV or profile must be judged against a role.
1 roleRead moreUse it when someone needs written feedback or a review cycle.
1 roleRead moreAccepting "done" as evidence of done. The status flip and the outcome are different events, and only one of them was checked.
A generalist agent absorbing specialist work, which makes it invisible on the specialist's board and produces a worse answer than the remit that was written for it.
Opening a role described by its title instead of by the work, so the interview loop measures nothing and the debate about the hire restarts at the offer.
Diagnosing a throughput drop as a motivation problem when the tracker shows the work is queued behind one reviewer.
Designing a loop as a sequence of conversations with senior people. An unstructured interview measures rapport and measures it repeatedly — five of them is one weak signal sampled five times, not five signals.
An onboarding plan made of reading and introductions, with no owned outcome — so nobody can tell whether it worked and the new person cannot either.
Producing a salary number from the model's own sense of the market. Compensation data is local, level-specific and dated; a confident figure with no named source is the single most damaging invented number in this domain, because somebody will offer it.
Scoring a candidate on impressions the CV produced — pedigree, company names, a tidy career line — rather than on evidence against a written bar. The impression is the part that carries bias and the part that predicts least.
Producing a review that grades a person rather than describing behaviour and its effect. A grade is arguable and forgettable; it also converts a performance conversation into a defence, which is the one thing that guarantees nothing changes.
Show us the work you want to delegate.
Book a tailored demo for AI Crew, Voice, Build or Missions. Track is available free for 30 days.