Pinned Loading

1. alignment-farm/materialize alignment-farm/materialize Public

   Turn a read manifest into an isolated, fail-closed workspace that contains exactly what's declared; hash-audited, and nothing else.

   Python 1

2. alignment-farm/construct alignment-farm/construct Public

   Research lab for agent-side memory: what becomes context, how prior experience changes later action, and when governance earns its cost.

   Python

3. alignment-farm/construct-2 alignment-farm/construct-2 Public

   How can model-backed agents improve from experience across sessions, and where should that experience live?

   HTML

4. alignment-farm/evidence-set-memory alignment-farm/evidence-set-memory Public

   Learns to assemble compact, sufficient evidence from prior work. Can an energy-based specialist can beat competent ordinary retrieval as needs change?

   Python

5. alignment-farm/experience-guided-investigation alignment-farm/experience-guided-investigation Public

   Can past debugging experience help an agent fix new problems better than just reading the code and history again?

   Python

6. alignment-farm/intervention-choice-transfer alignment-farm/intervention-choice-transfer Public

   When tools or interfaces change, can earlier repair outcomes teach an agent which fix to try next; better than diagnosing from scratch or reusing the old playbook?

   HTML