source_doc

active

source_doc:users-antonborzov-documents-research-nosync-alexander-applied-agent-harness-design-md

agent-harness-design.md

/Users/antonborzov/Documents/Research.nosync/alexander/applied/agent-harness-design.md

Frontmatter (2 fields)

{
  "path": "/Users/antonborzov/Documents/Research.nosync/alexander/applied/agent-harness-design.md",
  "last_claims_extracted_at": 1778899491
}

Outgoing (0)

None.

Incoming (24)

Source of (24)

A healthy agent ecology has a smooth trust gradient between user and agents, not a binary on/off.(claim)
A healthy agent ecology has rhythm—automated steps alternating with deferral and action alternating with review.(claim)
A robust agent ecology requires both automated-aligned action and explicit deferral—contrast between these regimes is essential.(claim)
A skill should do its work and stop, avoiding performative complexity and multi-stage flows when single-pass suffices.(claim)
A skill works at multiple reinforcing scales: individual command, skill invocation, and user project.(claim)
A skill's output length should match task complexity—short tasks produce short reports, long tasks produce structured hierarchical reports.(claim)
Anton has already built infrastructure that operationalizes roughly half of Alexander's 15 properties without explicit naming.(claim)
Anton's distinctive contribution would be care-centered and agent-ecology-centered, differentiating from Stefan Lesser's Alexander-software work.(claim)
Consistent voice, naming conventions, and hook patterns across skills are echo patterns in Alexander's sense.(claim)
Current application of 15 properties to agent harness is metaphorical; operationalization into measurement is an open empirical question.(claim)
Defining 'life' for agent ecologies requires explicit claim about adaptive responsiveness, compositional richness, or stewardship affordance.(claim)
Every skill should have one unmistakable thing it does; sprawling skills violate the Strong Centers principle.(claim)
Not-Separateness and Strong Centers can coexist in agent ecology, though this transfers from Alexander only by empirical test.(claim)
Not-Separateness in agent ecology means skills connect to user project, other skills, user history, and world data.(claim)
Parallel sub-tasks within skills and across skill families should produce parallel outputs for legibility.(claim)
+9 more