corpus-snapshot

Corpus Snapshot

Generate a point-in-time snapshot of the research corpus with computed metrics and analysis. Reads a snapshot template, fills [COMPUTE] sections with data, assists with [ANALYZE] sections, and writes the completed report.

Use the native CLI first:

aiwg corpus snapshot [--compute-only] [--delta-only] [--template <path>] \
  [--format full|summary|json] [--out <path>] [--date <YYYY-MM-DD>] [--write]

The command is backed by the declarative Flow playbook at agentic/code/frameworks/research-complete/flows/corpus-snapshot.playbook.yaml. The skill remains the orchestration wrapper for narrative [ANALYZE] sections.

Triggers

• "take a corpus snapshot"
• "generate corpus report"
• "snapshot the research"
• "corpus snapshot"
• /corpus-snapshot

Parameters

--compute-only (optional)

Only compute data sections — skip analysis sections. Faster, fully automated.

--delta-only (optional)

Only compute the delta from the previous snapshot. Useful for tracking session progress.

--template <path> (optional)

Custom template path. Default: .aiwg/reports/corpus-snapshot-template.md.

--format (optional)

Output format: full (default for the report file), summary (terminal), json (programmatic).

Prerequisites

Before generating a snapshot, the following should be current:

If prerequisites are stale, the snapshot will include warnings.

Execution Flow

Phase 1: Collect Raw Metrics

Scan the corpus and compute:

Dimensions:

• Total papers (node count)
• Total citation edges (edge count)
• Topics (unique tag count)
• Authors (unique author count)
• Year range (oldest → newest)
• Source types distribution

Topology (from citation-network index):

• Graph density: edges / (nodes * (nodes-1))
• Average degree (mean edges per node)
• Max hub (node with most connections)
• Connected components count
• Isolated nodes (degree 0)
• Diameter estimate (longest shortest path in largest component)

Degree Distribution:

• Histogram: how many nodes have degree 0, 1-2, 3-5, 6-10, 11-20, 20+
• Power law fit (if applicable)

Quality Distribution:

• GRADE breakdown: High / Moderate / Low / Very Low
• Doc depth: Full / Adequate / Stub / Skeleton (from quality-audit)
• Source availability: PDF present / Full text extracted / Missing

Phase 2: Compute Delta (if previous snapshot exists)

Compare current metrics against the most recent snapshot:

Delta from previous snapshot (2026-04-10):
  Papers:     +12 (360 → 372)
  Edges:      +87 (1,160 → 1,247)
  Density:    +0.001 (0.008 → 0.009)
  New topics:  +2 (gui-agents, code-generation)
  Stubs fixed: 23 (88 → 65)
  New hubs:    REF-364 (entered top 10)

Phase 3: Fill Template Sections

Read the snapshot template and fill sections:

[COMPUTE] sections — fully automated:

• Dimensions table
• Topology metrics
• Degree distribution histogram
• GRADE distribution
• Delta table

[ANALYZE] sections — agent-assisted:

• Cluster narrative: describe the main clusters and their themes
• Chain analysis: identify citation chains (A→B→C→D) and their significance
• Gap narrative: summarize disconnected areas and bridge opportunities
• Trend analysis: what's growing, what's stagnant

Phase 4: Write Report

Write the completed snapshot to:

.aiwg/reports/corpus-snapshot-YYYY-MM-DD.md

With frontmatter:

---
type: corpus-snapshot
date: 2026-04-13
papers: 372
edges: 1247
density: 0.009
components: 9
stub_rate: 0.17
previous: corpus-snapshot-2026-04-10.md
---

Phase 5: Report Summary

Corpus Snapshot Generated
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Papers: 372 (+12)  |  Edges: 1,247 (+87)
Density: 0.009     |  Components: 9
Hub: REF-016 (34)  |  Isolated: 3
GRADE: 33% High, 24% Mod, 26% Low, 16% VLow
Stubs: 65 (17%)    |  Full text: 54%

Delta highlights:
  +12 papers inducted
  +87 citation edges (backfill)
  -23 stubs (expanded)
  +2 new topics

Written to: .aiwg/reports/corpus-snapshot-2026-04-13.md

Template Format

The default template uses markers for computed vs analyzed sections:

# Corpus Snapshot — [DATE]

## Dimensions
[COMPUTE: dimensions-table]

## Topology
[COMPUTE: topology-metrics]

## Degree Distribution
[COMPUTE: degree-histogram]

## Quality Distribution
[COMPUTE: grade-distribution]
[COMPUTE: depth-distribution]

## Delta
[COMPUTE: delta-from-previous]

## Cluster Analysis
[ANALYZE: describe main clusters, their themes, and notable papers]

## Citation Chains
[ANALYZE: identify significant citation chains and their meaning]

## Gaps and Opportunities
[ANALYZE: summarize disconnected areas and bridge opportunities]

## Recommendations
[ANALYZE: what should be inducted next, what needs expansion]

Integration Points

Fortemi Core Migration Note

During the Fortemi Core index migration preview, corpus snapshots remain AIWG-rendered from corpus sidecars, corpus views, and the local .aiwg/.index artifacts. Do not use --backend fortemi-core as the source of truth for snapshot metrics until #1690 explicitly accepts a Fortemi-projected snapshot contract and #1691 parity fixtures prove identical REF/PROF, citation, profile, radar, discovery, and quality metrics.

Examples

# Full snapshot
aiwg corpus snapshot --write

# Just data, no analysis sections
aiwg corpus snapshot --compute-only --write

# Delta from previous snapshot only
aiwg corpus snapshot --delta-only

# Custom template
aiwg corpus snapshot --template .aiwg/reports/custom-template.md --write

# JSON metrics for dashboards
aiwg corpus snapshot --format json

References

• @$AIWG_ROOT/agentic/code/frameworks/research-complete/skills/corpus-index-build/SKILL.md — Index metrics source
• @$AIWG_ROOT/agentic/code/frameworks/research-complete/skills/research-quality-audit/SKILL.md — Depth distribution source
• @$AIWG_ROOT/agentic/code/frameworks/research-complete/skills/citation-backfill/SKILL.md — Prerequisite for topology
• @$AIWG_ROOT/agentic/code/frameworks/research-complete/skills/research-gap-detect/SKILL.md — Cluster data for narrative
• @$AIWG_ROOT/agentic/code/frameworks/research-complete/skills/research-status/SKILL.md — Health scoring complement