Skip to content
Skills Library
Data & analytics · Analysis

analyze

/analyze

Use when synthesising across prose and code corpora or ranking candidates by a criterion — tries the analyze plan library first (search → reference-chain traversal → rank → generate), inline-plans on a miss.

Repository
nexus ↗
Licence
AGPL-3.0
Last repository change
26.10.2026

Assessment and security

100/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content35 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
1
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 26.10.2026, without references to retired models or APIs.

Licence

AGPL-3.0, copyleft: przeróbki udostępniane innym muszą mieć tę samą licencję.

What it does and when to use it

Use when synthesising across prose and code corpora or ranking candidates by a criterion — tries the analyze plan library first (search → reference-chain traversal → rank → generate), inline-plans on a miss.

Original author description

Use when synthesising across prose and code corpora or ranking candidates by a criterion — tries the analyze plan library first (search → reference-chain traversal → rank → generate), inline-plans on a miss.

How it works

  1. The call
  2. Required bindings
  3. Optional bindings
  4. Typical intent shapes
  5. When direct search is fine
  6. Anti-patterns (do not do any of these)
Original SKILL.md headings · 68 lines, 408 words
SKILL.md preview
---
name: analyze
description: Use when synthesising across prose and code corpora or ranking candidates by a criterion — tries the analyze plan library first (search → reference-chain traversal → rank → generate), inline-plans on a miss.
effort: medium
---

**Tier-aware discipline** — before starting, check T3 (`search`), T2 (`memory_search`), and T1 (`scratch` search) widest to narrowest so you don't duplicate work already done; reuse a matching plan via `plan_search` before dispatching multiple agents. Before returning, write findings back at the tier matching their audience (`scratch` for siblings this session, `memory_put` for this project, `store_put` for permanent cross-project knowledge). Full checklist: [resources/tier-discipline.md](../../resources/tier-discipline.md) (shared across every skill that prescribes it — nexus-cnzei.6).

# analyze

**Call `nx_answer` for cross-corpus synthesis or ranking that must be
reduced from many documents** — the search → extract → rank → generate
composition the analyze plans provide. A single-corpus lookup is a
`search` call — see "When direct `search` is fine" below. This is the
one verb where bundling gives the biggest wins:
a 3-op chain collapses from ~45s of per-step spawns to ~15s in a single
`claude -p` call.

## The call

```
mcp__plugin_conexus_nexus__nx_answer(
    question=<caller's phrasing>,
    dimensions={"verb": "analyze"},
    context=<area, criterion, limit — as JSON string if needed>,
)
```

One tool call. `nx_answer` handles match → run → record, including the
operator-bundle optimization. Plan-miss falls through to an inline
`claude -p` planner.

## Required bindings

- `area` — the subject area being analysed.
- `criterion` — the axis used to rank candidates (e.g. "recency",
  "citation count", "coverage breadth").

## Optional bindings

- `limit` — per-corpus result cap (`default_bindings` → 12).

## Typical intent shapes

- "compare how X is handled across ml-systems and networking"
- "survey approaches to Y"
- "rank options for Z by cost"

## When direct `search` is fine

If the question is a single-corpus lookup — e.g. "find the chunks that
mention algorithm X" — use `mcp__plugin_conexus_nexus__search`. Analyze
earns its latency cost when the question requires multi-corpus
alignment, ranking by a semantic criterion, or structured extraction
before synthesis.

## Anti-patterns (do not do any of these)

- **Calling `search` directly for a cross-corpus synthesis question.**
  You get top-K chunks with no composition, no cross-corpus alignment,
  no ranking. If the question requires ranking or comparing across
  multiple collections, you need `nx_answer`'s full DAG.
- **Calling `plan_match` directly instead of `nx_answer`.** You lose
  the record step, the inline-planner fallback, and use_count telemetry.
- **Using `analyze` when `research` would suffice.** `research` is
  single-concept; `analyze` implies cross-corpus / cross-approach
  synthesis. If the caller only wants to understand one thing, use
  `/conexus:research-synthesis`.
- **Using bare `criterion`.** "important" is not a criterion; pick
  an axis the ranker can actually sort by.

See `/conexus:plan-first` and `docs/plan-authoring-guide.md`.

Versions

RepositoryScoreLicence
This page · hellblazer/nexus100AGPL-3.0
ericrisco/rsc-harness100MIT
rjmurillo/ai-agents100MIT
rjmurillo/ai-agents100MIT
rossumai/claude-marketplace100MIT
emaballarin/ccplugins100MIT
ramboz/jig100MIT
ramboz/jig100MIT
ramboz/jig100MIT

Database snapshot: 13.11.2026 · package version 7c8960b15c9b (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 100document
    Use when authoring or auditing documentation against existing coverage — tries the document plan library first (search prose/code → documentation-for traversal → compare), inline-plans on a miss.

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.