AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
TylerJNewman avatar

TylerJNewman

11 listings · 0 installs

Open-source publisher. Listings imported from github.com/TylerJNewman — credited to the original author with their license.

↗ github.com/TylerJNewman
11 results
Self-run
SKILL

Frame Loop

Frame one gated optimization target as a runnable autoresearch loop — the evaluation contract, a frozen harness, a loop mode with stop conditions, and a goal contract the agent cannot redefine. Use when a candidate is green and ready to loop, when scout-targets hands off a target, or when deciding whether a loop is safe to run overnight.

0
31
Free
Self-run
SKILL

Daily Codex Audit

Review recent Codex/agent work from receipts and produce a daily audit: what failed, what worked, what evidence supports it, and which process changes remain candidates. Use for daily or retro review of Codex sessions, logs, notes, tests, diffs, and local receipts. Do not use for greenfield loop design, automation authoring, or multi-day process promotion.

0
33
Free
Self-run
SKILL

Gates

Gate an optimization target against the seven properties that decide whether it is autoresearch-shaped — automatable, cheap, quiet, bounded, causal, un-gameable, navigable — plus a Goodhart pre-mortem. Use whenever scout-targets or frame-loop must check a target, or when the user asks whether something is worth optimizing with an automated loop, is autoresearch-shaped, or has an honest metric. Re…

0
46
Free
Self-run
SKILL

Design Deep Interfaces

Deepen human-facing interfaces by reducing decision load while preserving operational truth, control, and recovery. Use when designing or reviewing web/HTML, CLI, forms, dashboards, workflows, or conversational interfaces; or when improving defaults, disclosure, attention, accessibility, and safety.

0
29
Free
Self-run
SKILL

Johnny Decimal

Organize files, notes, projects, or any information with Johnny.Decimal (areas to categories to AC.ID, plus an index). Use when the user wants to design an organization system for folders or notes, asks where a file or item should be saved, needs a new ID or index (JDex) entry, asks to tidy or audit a messy folder tree, or has outgrown a system with more than 10 of something, needs more depth, or…

0
41
Free
Self-run
SKILL

Action Preserving Compression

Turn tacit experience into the coarsest heuristic that preserves competent action. Use when distilling conversations, transcripts, audits, examples, failures, or recurring decisions into handles; finding counterexamples; splitting overbroad heuristics; reviewing heuristic drift; or updating the heuristic catalog.

0
46
Free
Self-run
SKILL

Distill Law

Distill a system into a falsifiable law: one boundary, one conserved quantity, one probe, and one forced decision.

0
48
Free
Self-run
SKILL

Write Laws

Write falsifiable laws into the Basic Memory law project. Use when Codex needs to create, update, or teach law notes/lenses/coding principles in the law memory project, especially from transcripts, Oracle reviews, distill-law outputs, source documents, or conversation receipts; requires provenance, original source paths, source anchors, examples, index.md updates, and append-only log.md entries.

0
31
Free
Self-run
SKILL

Grug

>

0
39
Free
Self-run
SKILL

Judge Panel

Build and calibrate a multi-judge evaluator, then prove it agrees with held-out human labels before trusting it as a JUDGE. Use when an optimization target needs human-like judgment, when promoting an amber candidate toward a loop, or when another skill needs a trusted scoring panel.

0
41
Free
Self-run
SKILL

Scout Targets

Scout a codebase and work logs for autoresearch targets — places where a cheap, hard-to-game objective metric makes an automated optimization loop worth running. Use when the user wants to find where to apply autoresearch, nominate metric candidates, audit what is worth optimizing, or hunt optimization opportunities. Reach for it when another skill needs a ranked target ledger.

0
45
Free
You've reached the end · 11 loaded