msa-snapshot

Files (1)

• SKILL.md

SKILL.md (first lines)

# MSA Snapshot

A reference skill for the M3-C `output_format: table` mode, tuned for master services agreement portfolios. Produces a side-by-side grid of MSA-specific commercial terms across N agreements — the in-house lawyer's "compare these vendor MSAs" or "diligence these target MSAs" workflow. Each cell carries a citation back to the source document; failed extractions render as `not found` rather than confidently-wrong text.

## When this skill applies

Apply when the user has a portfolio of MSAs and wants to see how key commercial terms compare across them. Examples:

- "Pull liability caps, indemnification, and payment terms across these 8 vendor MSAs before our renewal cycle."
- "For this acquisition diligence, show me the renewal trigger and termination posture across the target's top 15 MSAs."
- "Compare the liability carveouts across our SaaS MSAs vs. our commercial-purchase MSAs."

Do not apply this skill to:

- Single-MSA review — use `msa-review-saas` or `msa-review-commercial-purchase` for one document at a time.
- General contract comparison across mixed types — use `contract-snapshot` for the general Term/Survival/Carveouts/Governing-Law grid.
- NDA portfolios — use `nda-snapshot` for NDA-specific columns (Confidential Information definition, permitted recipients, etc.).

## Pairing with the synthetic corpus

This skill ships paired with the synthetic MSA corpus in `docs/quickstart/sample-msas/` (5 MSAs with varying commercial terms). Operators trying LQ.AI for the first time can attach those 5 PDFs to a Knowledge Base and run this skill to see the tabular workflow end-to-end without committing real documents to the system.

## Fork-and-tune notes

The four columns here cover the highest-frequency MSA comparison questions for in-house counsel doing portfolio review or diligence. They do not overlap with the general `contract-snapshot` columns (Term, Survival, Carveouts, Governing Law) — operators wanting both can run the two skills in sequence.

When forking for your own MSA template / counterparty patterns, common modifications include:

- Adding a **Termination for Convenience** column if your business cares about exit flexibility (notice period + fee structure).
- Adding a **SLA / Service Levels** column for SaaS-heavy MSA portfolios (credit structure + measurement period).
- Adding a **Data Processing** column if you need to compare DPA references and data-residency commitments across vendors.
- Replacing the **Indemnification** column with a narrower **IP Indemnification** column if that is the only indemnification scope you care about.
- Bumping `minimum_inference_tier` to 3 on all columns for high-stakes diligence work.

The **Limitation of Liability** and **Indemnification** columns default to `minimum_inference_tier: 3` because these clauses are dense, fragmented across the document, and most prone to silent extraction errors. The carveouts in particular are the most-negotiated piece of an MSA — surfacing them inaccurately is worse than surfacing them not-at-all.

## Output expectations

For each document × column cell:

- A quoted phrase or short paragraph from the source document, anchored by character offsets to enable the citation modal.
- A brief plain-language summary when the operative clause spans multiple paragraphs.
- An explicit "one-way (vendor-favorable)" / "one-way (customer-favorable)" / "mutual" tag where the column asks about directionality (e.g., Indemnification).
- `not found` when the requested term is genuinely absent from the document (not when extraction failed — those surface as a parse error in the cell footer).

From LegalQuants/lq-ai, Apache-2.0.

More from LegalQuants/lq-ai