Co-Pilot
Updated a month ago

agent-scripts

Ssteipete
1.4k
steipete/agent-scripts
68
Agent Score

💡 Summary

A collection of reusable guardrail helpers for project synchronization and documentation.

🎯 Target Audience

Software developers managing multiple repositoriesProject managers overseeing documentation consistencyOnboarding specialists for new team membersDevOps engineers automating workflowsTechnical writers ensuring documentation quality

🤖 AI Roast:Powerful, but the setup might scare off the impatient.

Security AnalysisMedium Risk

The scripts may execute shell commands and interact with the filesystem, posing risks of command injection or unauthorized access. Ensure proper input validation and restrict execution contexts.

Agent Scripts

This folder collects the Sweetistics guardrail helpers so they are easy to reuse in other repos or share during onboarding. Everything here is copied verbatim from /Users/steipete/Projects/sweetistics on 2025-11-08 unless otherwise noted.

Additional skills (copied 2025-12-31) are from @Dimillian’s public Dimillian/Skills repository:

  • skills/swift-concurrency-expert
  • skills/swiftui-liquid-glass
  • skills/swiftui-performance-audit
  • skills/swiftui-view-refactor

Syncing With Other Repos

  • Treat this repo as the canonical mirror for the shared guardrail helpers. Whenever you edit scripts/committer or scripts/docs-list.ts in any repo, copy the change here and then back out to every other repo that carries the same helpers so they stay byte-identical.
  • When someone says “sync agent scripts,” pull the latest changes here, ensure downstream repos have the pointer-style AGENTS.MD, copy any helper updates into place, and reconcile differences before moving on.
  • Keep every file dependency-free and portable: the scripts must run in isolation across repos. Do not add tsconfig path aliases, shared source folders, or any other Sweetistics-specific imports—inline tiny helpers or duplicate the minimum code needed so the mirror stays self-contained.

Pointer-Style AGENTS

  • Shared guardrail text now lives only inside this repo: AGENTS.MD (shared rules + tool list).
  • Every consuming repo’s AGENTS.MD is reduced to the pointer line READ ~/Projects/agent-scripts/AGENTS.MD BEFORE ANYTHING (skip if missing). Place repo-specific rules after that line if they’re truly needed.
  • Do not copy the [shared] or <tools> blocks into other repos anymore. Instead, keep this repo updated and have downstream workspaces re-read AGENTS.MD when starting work.
  • When updating the shared instructions, edit agent-scripts/AGENTS.MD, mirror the change into ~/AGENTS.MD (Codex global), and let downstream repos continue referencing the pointer.

Committer Helper (scripts/committer)

  • What it is: Bash helper that stages exactly the files you list, enforces non-empty commit messages, and creates the commit.

Docs Lister (scripts/docs-list.ts)

  • What it is: tsx script that walks docs/, enforces front-matter (summary, read_when), and prints the summaries surfaced by pnpm run docs:list. Other repos can wire the same command into their onboarding flow.
  • Binary build: bin/docs-list is the compiled Bun CLI; regenerate it after editing scripts/docs-list.ts via bun build scripts/docs-list.ts --compile --outfile bin/docs-list.

Browser Tools (bin/browser-tools)

  • What it is: A standalone Chrome helper inspired by Mario Zechner’s “What if you don’t need MCP?” article. It launches/inspects DevTools-enabled Chrome profiles, pastes prompts, captures screenshots, and kills stray helper processes without needing the full Oracle CLI.
  • Usage: Prefer the compiled binary: bin/browser-tools --help. Common commands include start --profile, nav <url>, eval '<js>', screenshot, search --content "<query>", content <url>, inspect, and kill --all --force.
  • Rebuilding: The binary is not tracked in git. Re-generate it with bun build scripts/browser-tools.ts --compile --target bun --outfile bin/browser-tools (requires Bun) and leave transient node_modules/package.json out of the repo.
  • Portability: The tool has zero repo-specific imports. Copy the script or the binary into other automation projects as needed and keep this copy in sync with downstream forks. It detects Chrome sessions launched via --remote-debugging-port and --remote-debugging-pipe, so list/kill works for both styles.

Sync Expectations

  • This repository is the canonical mirror for the guardrail helpers used in mcporter and other Sweetistics projects. Whenever you edit scripts/committer, scripts/docs-list.ts, or related guardrail files in another repo, copy the changes back here immediately (and vice versa) so the code stays byte-identical.
  • When someone asks to “sync agent scripts,” update this repo, compare it against the active project, and reconcile differences in both directions before continuing.

@steipete Agent Instructions (pointer workflow)

  • The only full copies of the guardrails are agent-scripts/AGENTS.MD and ~/AGENTS.MD. Downstream repos should contain the pointer line plus any repo-local additions.
  • During a sync sweep: pull latest agent-scripts, ensure each target repo’s AGENTS.MD contains the pointer line at the top, append any repo-local notes beneath it, and update the helper scripts as needed.
  • If a repo needs custom instructions, clearly separate them from the pointer so future sweeps don’t overwrite local content.
  • For submodules (Peekaboo/*), repeat the pointer check inside each subrepo, push those changes, then bump submodule SHAs in the parent repo.
  • Skip experimental repos (e.g., poltergeist-pitui) unless explicitly requested.
5-Dim Analysis
Clarity8/10
Novelty5/10
Utility7/10
Completeness6/10
Maintainability8/10
Pros & Cons

Pros

  • Facilitates synchronization across multiple repos
  • Encourages consistent documentation practices
  • Portable and dependency-free scripts
  • Supports onboarding with shared resources

Cons

  • Limited to specific use cases
  • Requires manual syncing effort
  • Potential for outdated references if not maintained
  • May not cover all edge cases in documentation

Related Skills

pytorch

S
toolCode Lib
92/ 100

“It's the Swiss Army knife of deep learning, but good luck figuring out which of the 47 installation methods is the one that won't break your system.”

agno

S
toolCode Lib
90/ 100

“It promises to be the Kubernetes for agents, but let's see if developers have the patience to learn yet another orchestration layer.”

nuxt-skills

S
toolCo-Pilot
90/ 100

“It's essentially a well-organized cheat sheet that turns your AI assistant into a Nuxt framework parrot.”

Disclaimer: This content is sourced from GitHub open source projects for display and rating purposes only.

Copyright belongs to the original author steipete.