Packs
Results
Adversarial Change Review
Find the defects in a code change you did not write, and say nothing about style.
Behaviour-First Test Design
Get tests that fail when the code is wrong, and prove that they do.
Contract Review Against a Playbook
Review a contract against your own positions, and be told what you did not find as well as what you did.
Decision Memo
Turn an open decision into a memo that gets the decision made and recorded.
Evidence-Based Market Research
Find out what is actually known about a market, with every figure traceable and every gap named.
Verified Literature Review
Produce a literature review whose every citation exists and says what you claim it says.
6 built. 21 specified and coming. 4 killed — and we say why.
Every pack below already exists as a specification: a job statement, the phrases that trigger it, and the test cases it has to survive. Building one is the last step, not the first — which is why this list is worth publishing before the shelf is full.
6 of 27 shippable · 22%
Specified and queued. Each already has its job statement, triggers and test cases written.
- capability engineering
Capability Authoring & Hardening
Write a capability that activates when it should and keeps working as it grows.
In the queue - engineering
Codebase Orientation
Produce a durable orientation artifact for a codebase nobody documented.
In the queue - strategy
Customer Interview Synthesis
Turn customer conversations into decisions, separating what people say from what they do.
In the queue - communication
Executive Storyline
Turn source material into an assertion-first storyline that survives a partner reading.
In the queue - communication
Meeting to Decisions & Commitments
Turn a meeting transcript into decisions and commitments with named owners.
In the queue - capability engineering
Output Evaluation Rubric
Build a reusable rubric for a class of LLM output, and a procedure for applying it.
In the queue - operations
Recurring Operating Report
Produce the recurring operating report without redoing the assembly every time.
In the queue - engineering
Refactor & Migration Plan
Plan a large change to working code as verifiable stages with an abort point at each.
In the queue - operations
Spec to Engineering Handoff
Turn an idea into a spec an engineer can build from without asking a question.
In the queue - research
Systematic Review Screening
Screen a large candidate set against pre-registered criteria, defensibly.
In the queue - finance
Traceable Financial Analysis
Produce financial analysis where every figure walks back to a source cell.
In the queue
Specified, further out. Sequenced behind Tier 2 because the research is thinner.
- capability engineering
Agent Constraint Design
Stop an agent looping, drifting or exceeding its brief.
In the queue - capability engineering
Connector Selection & Verification
Decide which MCP connectors earn the flakiness they add.
In the queue - capability engineering
Context Budget Design
Decide what to load and what to exclude for a long task.
In the queue - operations
Data Room Readiness
Assemble and gap-check a diligence data room.
In the queue - engineering
Incident Investigation
Trace a production failure across services from logs and traces.
In the queue - communication
Investor Update Assembly
Assemble the monthly investor update; the bottleneck is gathering, not writing.
In the queue - research
Method-Level Paper Comprehension
Read one paper to the depth where you could critique its method.
In the queue - communication
Performance Review Evidence
Assemble evidence for a review; the manager still writes it.
In the queue - communication
RFP / Proposal Response
Answer an RFP from real past work rather than from imagination.
In the queue - operations
Support Knowledge Base Refresh
Regenerate a knowledge base from recent tickets so it stops going stale.
In the queue
Specified, researched, then stopped. These stay on the list with the reason attached — a pack that quietly disappears teaches nobody anything.
- strategy
Idea Red-Team
Why it stopped: Pressure-test an idea adversarially. NOT RECOMMENDED as a standalone pack -- see spec.
Not being built - research
Grant Proposal Drafting
Why it stopped: too funder-specific to generalise without a real funder corpus.
Not being built - strategy
Pricing Decision
Why it stopped: no repeatable method found in the research.
Not being built - strategy
Roadmap Prioritisation
Why it stopped: the research describes organisational politics, not a procedure.
Not being built
If one of these is the job you actually have, that is the useful thing to tell us — it is how the queue gets reordered.
Find what is built for your job