Skip to content

aicraftbook/aicraftbook-skills / skills/prompt-proof

prompt-proof

Verified

Reviews a system prompt or prompt template before it goes into an app or automation. Finds ambiguous instructions, conflicting rules, missing output rules, places where interpolated user text could override the instructions (prompt injection), and missing handling for empty or unexpected input, then returns the issues ranked by impact, a revised prompt that keeps the author's intent, and test inputs with what a correct response must and must not do. Use when someone shares a system prompt, developer message, prompt template with variables such as {{input}}, or an LLM step from n8n, Make, Zapier or code and asks "review my prompt", "what could go wrong with this prompt", "is this prompt safe", "make this prompt more reliable" or "write tests for this prompt".

FileKindSize
CHANGELOG.mdmarkdown1.6 KB
LICENSEother1.0 KB
README.mdmarkdown11.7 KB
SKILL.mdskill md8.1 KB
evals/independent-2026-10-21.jsonconfig66.6 KB
evals/independent-2026-10-22-b.jsonconfig57.1 KB
evals/independent-2026-10-22.jsonconfig58.0 KB
evals/results-2026-10-22.jsonconfig5.3 KB
evals/tasks.jsonconfig7.1 KB
examples/requests.mdmarkdown1.5 KB
references/failure-modes.mdmarkdown11.7 KB
references/test-template.mdmarkdown4.5 KB
skill.jsonconfig1.5 KB

CHANGELOG.md

View on GitHub
# Changelog

## 1.1.0 (2026-10-22)

- New rules: quote prompt text and code exactly, check the order of sections before describing it, report nothing as present that cannot be quoted, and do not state how a provider ranks message roles from memory.
- New final step: check every issue against the original prompt before answering. Tests are labelled proposed (not run).
- A second independent evaluation (2026-10-22) found factual slips in 8 of 12 runs with 1.0.2; on the same frozen tasks 1.1.0 had 2 of 12 (4 of 12 without the skill).

## 1.0.2 (2026-10-22)

- Documentation only: README rewritten as a full reference (install per agent, permissions, examples, troubleshooting, test results). SKILL.md and references unchanged.

## 1.0.1 (2026-10-21)

- `references/failure-modes.md` §4 (injection): the "After" example no longer removes the author's
  `issue_refund` tool; it keeps the tool behind human approval, says that dropping it would be a change of
  intent, and adds stripping or escaping the closing tag (or JSON-encoding the value) before the customer
  text is inserted.
- `references/test-template.md`: the example prompt now states its priority rule for messages that fit
  more than one category, and test T4 checks for the category that rule gives.
- The evaluations ran on 1.0.0.

## 1.0.0 (2026-10-21)

- First draft. Six failure modes (ambiguity, conflicting rules, missing output contract, injection
  exposure, empty and unexpected input, over-long instructions) with before/after examples, impact levels,
  a fixed output format, and a test case template. Guidance checked against Anthropic, OpenAI and OWASP
  sources on 2026-10-21. Not yet evaluated.
finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the workStars: 323kUpdated Claude CodeCodexVerifiedInstall
systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixesStars: 323kUpdated Claude CodeCodexVerifiedInstall
test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation codeStars: 323kUpdated Claude CodeCodexVerifiedInstall
skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.Stars: 198kUpdated Claude CodeCodexVerifiedInstall