fai-tldr-prompt — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited fai-tldr-prompt (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Generate concise, actionable summaries from long content.
def summarize_document(text: str, max_bullets: int = 5) -> str:
return llm(f"""Summarize in {max_bullets} bullet points.
Rules:
- Each bullet is one clear sentence
- Focus on decisions, actions, and key facts
- Skip background and filler
- Include numbers and specifics where available
Document:
{text}""")def summarize_pr_review(comments: list[dict]) -> str:
formatted = "\n".join(f"- [{c['severity']}] {c['file']}:{c['line']}: {c['comment']}"
for c in comments)
return llm(f"""Summarize this code review in 3-5 bullets.
Group by: blocking issues, improvements, positive feedback.
Comments:
{formatted}""")def summarize_thread(messages: list[dict]) -> str:
formatted = "\n".join(f"{m['author']}: {m['content']}" for m in messages)
return llm(f"""Summarize this discussion thread.
Output:
- Decision: [what was decided]
- Action items: [who does what by when]
- Open questions: [unresolved items]
Thread:
{formatted}""")| Check | Requirement |
|---|---|
| Length | 3-5 bullets for short docs, 5-10 for long |
| Specifics | Numbers, names, dates — not vague |
| Actionable | Reader knows what to do next |
| Accurate | No hallucinated details |
| Self-contained | Makes sense without reading original |
| Format | Template |
|---|---|
| Bullet | - Key point 1\n- Key point 2 |
| Executive | ## Summary\n[paragraph]\n## Actions\n[list] |
| One-liner | "[Subject] — [key decision] — [next step]" |
| Issue | Cause | Fix |
|---|---|---|
| Summary too vague | No specifics rule | Add "include numbers and names" |
| Important detail missed | Content too long | Chunk and summarize sections, then merge |
| Hallucinated facts | Model fills gaps | Add "only include facts from the document" |
| Too long | No length constraint | Set explicit bullet count limit |
| Practice | Rationale |
|---|---|
| Always include grounding rules | Prevents hallucination |
| Set explicit output format | Consistent, parseable responses |
| Use temperature 0-0.3 for factual tasks | Reduces randomness |
| Add few-shot examples for complex formats | Shows model the expected pattern |
| Test prompts with evaluation dataset | Measure quality objectively |
| Version control all prompts | Track changes, enable rollback |
fai-prompt-builder — Structured prompt constructionfai-boost-prompt — Interactive prompt refinementfai-basic-prompt-optimization — Prompt quality patternsfai-finalize-agent-prompt — Production prompt validation~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.