Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/xxyoudeadpunkxx/memory-assisted-shaping/llms.txt

Use this file to discover all available pages before exploring further.

The protocol defines explicit failure and success conditions. These are not aspirational goals — they are the mechanical tests of whether the protocol is operating correctly. If a failure condition is met, the session has gone wrong regardless of how the output looks.

Failure conditions

  • Builds before reading
  • Brainstorms by default
  • Diagnoses without evidence
  • Treats inference as evidence
  • Formalizes from a single signal
  • Asks about reversible non-gates instead of proposing defaults
  • Defers a gate and later treats it as closed
  • Mixes philosophy, constraints, runtime, architecture, artifact structure, and implementation detail
  • Imposes a fixed artifact family on every idea
  • Enters non-trivial synthesis with unresolved uncertainty about gates, decisions, discarded paths, or artifact readiness
  • Follows memory signals over OP current instruction
  • Treats memory signals as authority
  • Reads or rewrites live memory files during shaping
  • Treats a memory signal as confirmed structure
  • Emits memory signals for cosmetic or non-operational details
  • Turns memory signaling into a checklist
  • Claims persisted memory when Python persistence is unavailable
  • Outputs a memory signal in chat instead of appending it when append is required and Python is available
  • Produces artifacts before explicit OP approval
  • Contaminates final artifacts with process trace
  • Preserves flexibility at the cost of operability
  • Uses elegant language instead of mechanical clarity
  • Exposes provenance or evidence labels when they do not reduce ambiguity

Success conditions

The protocol succeeds when all of the following hold:
  • The idea becomes structurally legible
  • Active tensions expose useful friction or converge into simpler form
  • Real gates are few and explicit
  • Non-gates resolve without drag
  • OP current instruction remains dominant
  • Memory signals preserve continuity without becoming the work surface
  • Later reinterpretation can recover gates, decisions, discarded paths, shape movement, source boundaries, and artifact boundaries from minimal signals
  • Synthesis can produce a clean artifact without redesign or process leakage
If you’re unsure whether a session is going well, ask GPT to summarize the internal notes. The structure of the summary reveals protocol health — few gates, clean decisions, no zombie discarded paths. A summary cluttered with open tensions, unresolved gates, and competing shapes is a signal that the session needs more reading before it proceeds to shaping.

Build docs developers (and LLMs) love