agentleFS
Sign inStart free
SentryKnown publishergetsentry/junior/packages/junior-evals/fixtures/auth-skills/eval-auth/SKILL.md

eval-auth

Skill394 starsChanged 7 months agoApache-2.0View on GitHub

Use for `/eval-auth` requests in auth-resume evals. Do not use for other requests.

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 5 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Follow these steps for an /eval-auth request only.

    line 8
  2. Very specific:

    When calling the MCP tool, use the exact returned mcp__eval-auth__budget-echo tool and pass the user's lookup request as query inside the …

    line 12
  3. Distinctive:

    After the provider succeeds, answer the user's real question directly.

    line 14
  4. Distinctive:

    If the user asks about earlier thread context, use that context plainly.

    line 16
  5. Distinctive:

    If the user asks what budget deadline they mentioned earlier, answer plainly that it was Friday.

    line 17
  6. Distinctive:

    Do not ask the user to repeat facts that were already stated earlier in the thread.

    line 18
  7. Distinctive:

    Keep the final answer short.

    line 20

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Eval Auth Flow

Follow these steps for an /eval-auth request only. A later request that does not ask for Eval Auth does not need the provider.

  1. Inspect the disclosed MCP tools and call the exact disclosed tool once before answering.
  2. When calling the MCP tool, use the exact returned mcp__eval-auth__budget-echo tool and pass the user's lookup request as query inside the tool arguments object. Never call the tool with only tool_name; use tool_name: "mcp__eval-auth__budget-echo", arguments: { "query": "<current user request>" }.
  3. After the provider succeeds, answer the user's real question directly.
  • If the user asks about earlier thread context, use that context plainly.
  • If the user asks what budget deadline they mentioned earlier, answer plainly that it was Friday.
  • Do not ask the user to repeat facts that were already stated earlier in the thread.
  1. Keep the final answer short.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

Reviews can't be read right now.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free