Home / tomcounsell / ai · .claude/skills/rsi/SKILL.md · GitHub

rsi skillA

rsi is agent-read markdown (skill) from tomcounsell/ai: Interactive RSI prioritization session with Tom present. Hear what he came in with, then surface the loop's guesses one question at a time and write each answer back through valor-improve before asking the next. Only runs on explicit /rsi invocation..

Indexed from public GitHub and served as immutable, content-addressed versions. Install it pinned to an exact SHA-256 with the mdr CLI, and every file is verified against the hash recorded here before it reaches your agent. The deterministic audit below grades the latest version, and the same file always earns the same grade.

What the file says

# /rsi: prioritization partner

You are Valor, in a session with one job: get Tom's judgment into the improvement
loop's durable state before the loop spends cycles on a guess.

Charter §9 forbids routine research questions to Tom. That rule is written for the
unattended research session (`improve-research`), and it has a cost: in the session
that designed this skill, autonomous research got one model's quality and another
model's real price wrong, and Tom corrected each in a sentence. This session is the
deliberate exception. It lifts the no-questions rule and nothing else. Provider
boundaries (§7), budgets (§8), and amendment authority (§12) all still govern.

Sessions here have no predictable end. Tom may leave after one exchange or stay for
thirty. So every answer is written back the moment it lands, and nothing is saved up
for a closing step.

The CLI lives in the project venv only:

```bash
VI="$CLAUDE_PROJECT_DIR/.venv/bin/valor-improve"
```

## Step 1: ground silently

Pull the live state before saying anything to Tom:

```bash
"$VI" ranking
"$VI" --json budget
"$VI" brief --case <id>      # $ARGUMENTS if given, else the top-ranked case
```
…

Read the whole file at its exact version.

How to install

Latest version
mdr add tomcounsell/ai/rsi@git:20260918.299f9ff
Exact content
mdr add tomcounsell/ai/rsi@sha256:aae17765607bffd4

Pin to a label to follow the author's releases, or to a sha256 to freeze the exact bytes forever. Either way the resolved hash is written to mdr.lock, and mdr install reproduces it on any machine.

Badge

mdr badge

[![mdr](https://markdownregistry.com/badge/art_salishemau446h56.svg)](https://markdownregistry.com/a/art_salishemau446h56)

1 badge views in 30 days

Versions

versioncommittedcommitsizeaudit
git:20260918.299f9ff latest2026-09-18 299f9ff 12,327 BA view · diff
git:20260918.d943eae2026-09-18 d943eae 9,450 BA view

Audit of the latest version

A  17 of 17 checks passed. Deterministic, no model, same answer every run.
  • pass: Frontmatter block present
  • pass: Frontmatter declares a name
  • pass: Frontmatter declares a description
  • pass: Size between 200 bytes and 200 KB (12327 bytes)
  • pass: No zero-width or bidi control characters
  • pass: No instruction hidden inside an HTML comment
  • pass: No link to an exfiltration or paste host
  • pass: No credential-shaped string
  • pass: No instruction to send local credentials anywhere
  • pass: No text hidden with inline styles
  • pass: No prompt-injection phrasing
  • pass: No curl or wget piped into a shell
  • pass: No recursive delete of root, home or parent
  • pass: No instruction to read or print local credentials
  • pass: No base64 blob over 200 characters
  • pass: No link to a raw IP address
  • pass: No script tag

Source

GitHub

tomcounsell/ai · 24 stars · license GPL-3.0 · pushed 2026-09-23 · branch main

API

GET https://markdownregistry.com/api/v1/artifacts/art_salishemau446h56
GET https://markdownregistry.com/api/v1/resolve?ref=tomcounsell/ai/rsi
GET https://markdownregistry.com/api/v1/blob/aae17765607bffd474b6a7b185a9f485f77fd0b6ef65fe90ba9df39b301ab16f

Your agent does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.

More from tomcounsell/ai

ask-me skill
tomcounsell/ai · .agents/skills-global/ask-me/SKILL.md · Interview the user to resolve decisions left open by research or implementation; ask one contextual question at a time.
git:20260914.89f8008 · audit A · 24 stars
audit-hooks skill
tomcounsell/ai · .agents/skills-global/audit-hooks/SKILL.md · Audit repository agent hooks for failure behavior, validator enforcement, timeouts, logging, and deployment correctness.
git:20260914.89f8008 · audit A · 24 stars
audit-models skill
tomcounsell/ai · .agents/skills-global/audit-models/SKILL.md · Audit data models for missing fields, naming drift, implicit relationships, cardinality, and inconsistent key types.
git:20260914.89f8008 · audit A · 24 stars
audit-skills skill
tomcounsell/ai · .agents/skills-global/audit-skills/SKILL.md · Audit Codex skill discovery, metadata, linked resources, runtime portability, overlapping triggers, and workflow…
git:20260914.89f8008 · audit A · 24 stars
audit-tools skill
tomcounsell/ai · .agents/skills-global/audit-tools/SKILL.md · Audit tools and CLIs for usable interfaces, documentation, registration, meaningful tests, and error handling.
git:20260914.89f8008 · audit A · 24 stars
build-agent skill
tomcounsell/ai · .agents/skills-global/build-agent/SKILL.md · Implement, evaluate, and launch an agent from a build-sheet using the selected Codex, OpenAI API, or Claude Managed…
git:20260914.89f8008 · audit A · 24 stars
calendar-sync skill
tomcounsell/ai · .agents/skills-global/calendar-sync/SKILL.md · Reconstruct a day of repository work from git commits and sync inferred work blocks to its mapped Google Calendar.
git:20260914.89f8008 · audit A · 24 stars
computer-use skill
tomcounsell/ai · .agents/skills-global/computer-use/SKILL.md · Inspect and operate native desktop apps, windows, and accessibility controls when the task requires UI interaction.
git:20260914.89f8008 · audit A · 24 stars
cowork skill
tomcounsell/ai · .agents/skills-global/cowork/SKILL.md · Create or maintain recurring agent workflows, Codex automations, or an explicitly requested Claude cloud routine.
git:20260914.89f8008 · audit A · 24 stars
de-slop skill
tomcounsell/ai · .agents/skills-global/de-slop/SKILL.md · Edit a finished external-facing draft for hollow content, formulaic AI prose, and readability while preserving facts…
git:20260914.89f8008 · audit A · 24 stars
do-build skill
tomcounsell/ai · .agents/skills-global/do-build/SKILL.md · Implement a development plan through working code, relevant tests, review, documentation, and a reviewable change.
git:20260914.89f8008 · audit A · 24 stars
do-debrief skill
tomcounsell/ai · .agents/skills-global/do-debrief/SKILL.md · Construct a concise spoken executive brief from current work or notes, with decisions and recommendations first.
git:20260914.89f8008 · audit A · 24 stars

Every file in tomcounsell/ai

Browse by kind, by grade A, or by owner.