Automate growth-loop attribution and hypothesis confidence updates across growth signals.
npx skills add https://github.com/0xHoneyJar/construct-observer --skill growing
Please help me install this Agent Skill. Skill: growing Source: https://github.com/0xHoneyJar/construct-observer/tree/main/skills/growing Command: npx skills add https://github.com/0xHoneyJar/construct-observer --skill growing
Reviews and orchestrates the growth-loop lifecycle by validating proposed matches, updating hypothesis confidence, executing decay, and surfacing growth changes in a controlled, operator-verified pipeline.
Run the growth loop for all users with /grow or target a single user with /grow --user <user> and allow operator confirmations during the review steps.
High-intent search queries and answers about installing and using this skill.
You can orchestrate the full lifecycle by running /grow for all users or /grow --user <user> for a single target. The pipeline executes operator-confirmed reviews, updates hypothesis confidence, and applies decay to aging signals automatically.
Hypothesis confidence decay reduces the influence of aging growth signals over time. The Skill applies atomic state updates to aging hypotheses, ensuring outdated signals lose influence while operator-verified matches maintain attribution accuracy.
Yes, per-user isolation during classification preserves individual context and prevents cross-user leakage. This ensures deterministic and manual review workflows maintain strict boundaries when attributing growth signals across multi-user programs.
Operator-confirmed review validates proposed growth matches before applying state updates, ensuring signal quality is correctly attributed. This manual verification layer prevents false positives from contaminating hypothesis confidence and cross-user pattern detection.
Yes, cross-user pattern discovery surfaces systemic growth signals and decay trends across the cohort. It identifies shared growth changes while maintaining per-user isolation during the classification and operator-verified review cycles.
Manual review workflows require operator confirmation for proposed matches, which can create bottlenecks at scale. However, this constraint ensures signal quality attribution and prevents unvalidated hypotheses from affecting cross-user pattern detection and decay execution.
calcosmicIngest and normalize project documentation into a structured colony knowledge base.
0xHoneyJarGenerates per-user cognition sidecars with provenance from canvas data.
jaskaranhundalNormalizes raw security events into typed, attributed, confidence-scored SecurityFact objects.
nerdsaneDeploy formally verified agent operating systems with Cedar authorization.
