System Prompt Lookup

What it does

When your agent uses it

• The user asks what some agent is instructed to do
• Quotes a system prompt and wants it verified
• Shows an extracted
• Compares how two products are set up

Example prompts

• “Use the system-prompt-lookup skill to check what a shipped AI product's system prompt and tool schema actually say by reading a dated public archive…”
• “/system-prompt-lookup”

Workflow steps

5 steps, taken from the first numbered list in SKILL.md.

1. 1Finds the product's directory and reads its README.md, which lists each file with model, mode, character count and tool count.
2. 2Fetches the actual prompt and tool schema rather than summarising from memory.
3. 3Checks docs/CAPTURES.md to see whether the artifact was captured off the wire or reported by the vendor, and says which.
4. 4Diffs when more than one artifact exists for a product, because the differences are the finding.
5. 5States plainly when the archive has no entry for what was asked.

What it can do on your machine

Read from SKILL.md and the folder at commit 616deb5. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  Shell commands in SKILL.md call:

  • curl
  • claude

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  Hosts in commands or code, which the agent is likely to contact:

  • api.github.com
  • raw.githubusercontent.com

  Also links to:

  • github.com

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

System Prompt Lookup loads about 1.2k tokens when it runs. Until then it costs about 98 tokens; SKILL.md has 476 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~98

When it runs · the whole SKILL.md, loaded when a task matches

~1.2k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from davepoon/buildwithclaude at commit 616deb5, republished under its AGPL-3.0 licence (© davepoon). 476 words, ~1,185 tokens.

name

system-prompt-lookup

description

Checks what a shipped AI product's system prompt and tool schema actually say by reading a dated public archive instead of recalling them. Use when the user asks what some agent is instructed to do, quotes a system prompt and wants it verified, shows an extracted or leaked prompt, or compares how two products are set up. Read-only; fetches public files from github.com.

category

ai-agents

license

AGPL-3.0

requires.bins

curl

What does Claude Code's system prompt actually say about running tests?

# What products exist
curl -s https://api.github.com/repos/Continuum-AI-Corp/OrcaPromptVault/contents | grep '"name"'

# Read one artifact
BASE=https://raw.githubusercontent.com/Continuum-AI-Corp/OrcaPromptVault/main
curl -s "$BASE/Claude-Code/claude-code-opus-5-system-prompt-2026-09-03.md" | head -40
curl -s "$BASE/Claude-Code/claude-code-opus-5-tools.json" | grep -o '"name": *"[^"]*"'

# Check provenance before relying on it
curl -s "$BASE/docs/CAPTURES.md" | grep -i "claude code"

Yes, but not in the prompt most people mean.

Claude Code sends a different prompt depending on how it is started. The interactive
CLI capture opens "You are Claude Code, Anthropic's official CLI for Claude"; the
headless / Agent SDK capture opens with the line you quoted. Same model, same day.

  interactive  claude-code-fable-5.1-system-prompt-2026-09-02.md        26,131 chars, 35 tools
  headless     claude-code-fable-5.1-print-system-prompt-2026-09-02.md  20,806 chars, 29 tools

Both are wire captures per docs/CAPTURES.md, not vendor-published text. So the line
is genuine, and it tells you which entry point produced it.

© davepoon, AGPL-3.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in plugins/all-skills/skills/system-prompt-lookup of davepoon/buildwithclaude.

Open the folder on GitHubat commit 616deb5