Skip to content
Agentmart
Submit a tool

hm-l3-research-chain

Orchestrate the canonical research chain: ingest → detect → research → synthesize → artifact.

Skill by shynlee04

IndexedMIT★ 51
Write a review

Chalk Labs

Need help deploying hm-l3-research-chain?

Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.

Book a call

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/shynlee04-hm-l3-research-chain

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Orchestrate the canonical research chain: ingest → detect → research → synthesize → artifact. Use when starting a research task, chaining investigation skills, producing a final artifact from multiple sources, when the user needs comprehensive analysis, or when research needs to be structured and repeatable. Even when the user says "look into this" or "find out about." Triggers: "research", "investigate", "analysis", "deep dive", "comprehensive research", "research chain", "multi-source", "synthesize findings". Sibling skills: use with "hm-detective" for codebase detection, "hm-deep-research" for evidence gathering, "hm-synthesis" for compression, "hm-tech-stack-ingest" for dependency caching. NOT for single-source lookups or quick fact-checking.

Use cases: research, productivity

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md and code. Checked Nov 1, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found
Runs as
Shell scripts

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/hm-l3-research-chain

What's inside

Skill name
hm-l3-research-chain
Description
Orchestrate the canonical research chain: ingest → detect → research → synthesize → artifact. Use when starting a research task, chaining investigation skills, producing a final artifact from multiple sources, when the user needs comprehensive analysis, or when research needs to be structured and repeatable. Even when the user says "look into this" or "find out about." Triggers: "research", "investigate", "analysis", "deep dive", "comprehensive research", "research chain", "multi-source", "synthesize findings". Sibling skills: use with "hm-detective" for codebase detection, "hm-deep-research" for evidence gathering, "hm-synthesis" for compression, "hm-tech-stack-ingest" for dependency caching. NOT for single-source lookups or quick fact-checking.
License
MIT
Metadata
role: orchestrator · layer: 1 · pattern: P1 · version: 1.0.0
Instructions
228 lines in SKILL.md · view on GitHub

Before installing

  • Ships executable scripts

Agent Skills spec issues

  • allowed-tools: must be a space-separated string

Files (9)

  • SKILL.md
  • evals.json · evals/
  • .gitkeep · metrics/
  • rich-gate-scorecard.md · metrics/
  • chain-stages.md · references/
  • tool-matrix.md · references/
  • validate-skill.sh · scripts/
  • .gitkeep · templates/
  • chain-continuation.md · templates/

SKILL.md (first lines)

## Overview

Orchestrate the canonical research pipeline: detect → research → synthesize → artifact. Use when starting research tasks, chaining investigation skills, producing final artifacts from multiple sources, or when research needs to be structured and repeatable. Produces synthesized research outputs with citation tracking and evidence persistence.

## The Iron Law

```
Research without synthesis is hoarding. Synthesis without evidence is hallucination.
Ingestion without caching is repetition.
```

# Research Chain
## The Canonical Chain

```
┌──────────┐     ┌─────────────┐     ┌──────────────┐     ┌───────────┐     ┌──────────┐
│  INGEST  │ ──→ │   DETECT    │ ──→ │   RESEARCH   │ ──→ │ SYNTHESIZE│ ──→ │ ARTIFACT │
│(hm-tech-  │     │(hm-detective)│     │(hm-deep-     │     │(hm-synthesis)│   │ (export) │
│ stack-ingest)│  │             │     │   research)  │     │           │     │          │
└──────────┘     └─────────────┘     └──────────────┘     └───────────┘     └──────────┘
```

### Stage 0: Ingest (hm-tech-stack-ingest)

**Purpose:** Cache third-party repositories, SDK docs, and API references BEFORE researching. This ensures all downstream stages validate against REAL code, not assumptions.

```bash
# Check if the target tech stack is already cached
ls references/tech-stacks/<name>/ && cat references/tech-stacks/<name>/metadata.json

# If missing or stale, run the ingestion pipeline
# hm-tech-stack-ingest handles: DETECT → DISCOVER → INGEST → ORGANIZE → INDEX
```

**Output:** Cached tech stack in `references/tech-stacks/<name>/` with TOC.md, metadata.json, api/, docs/, examples/

**Gate:** Cached assets must have valid `metadata.json` with `version`, `source_url`, and `ingest_date`. If the cached version differs from the installed version, re-ingestion must complete before research starts.

### Stage 1: Detect (hm-detective)

**Purpose:** Understand what you are looking at before researching.

```bash
# SCAN mode — tech stack detection
# READ mode — targeted file reads
# DEEP mode — comprehensive codebase scan
```

**Output:** `.tech-registry.json` + initial findings

**Gate:** detection output must name scope, search modes used, and any missing/inaccessible areas before research starts.

### Stage 2: Research (hm-deep-research)

**Purpose:** Gather evidence using version-matched documentation and MCP tools.

```bash
# Context7 for API verification
# Tavily for current information
# Repomix for codebase structure
```

**Output:** Structured findings with citations

**Gate:** research output must include source evaluation, contradiction status, and blocked-source notes before synthesis starts.

### Stage 3: Synthesize (hm-synthesis)

**Purpose:** Compress findings into actionable artifacts.

```bash
# Compression tiers: Snapshot → Focused → Signature
# Artifact export: markdown, JSON, or plan files
```

**Output:** Final artifact (report, plan, or specification)

**Gate:** synthesis output must include methodology/limitations and link each recommendation to evidence.

### Stage 4: Artifact + Continuation

**Purpose:** Persist lineage so a later agent can resume without rediscovering the chain.

**Required artifact metadata:**

```yaml
research_chain_id: YYYY-MM-DD-topic-slug
detect_artifact: path-or-summary
research_artifact: path-or-summary
synthesis_artifact: path-or-summary
sources_reviewed: []
blocked_sources: []
contradictions: resolved | unresolved | none
next_action: verify | implement | ask | block
```

**Stop rule:** If a required stage artifact is missing, return `BLOCKED` with the missing gate. Do not synthesize or mark complete from partial chain state.

## When to Use the Full Chain

| Situation | Use Chain? |
|-----------|-----------|
| Exploring an unfamiliar codebase | Yes — full chain |
| Verifying a single API signature | No — Context7 only |
| Writing a project specification | Yes — full chain |
| Quick bug investigation | Partial — detect + research, skip synthesis |
| Producing a migration plan | Yes — full chain |

## Anti-Patterns

| Anti-Pattern | Detection | Correction |
|-------------|-----------|------------|
| **The Skipped Detect** | Jumps straight to research without scanning | Always run hm-detective SCAN first |
| **The Hoarder** | Gathers 50 sources but never synthesizes | Cap research time, force synthesis stage |
| **The Single-Source** | Uses only one tool (e.g., only web search) | Use the full MCP matrix |
| **The Orphan Artifact** | Produces artifact but never links to source evidence | Every claim in artifact must cite source |
| **The Ungated Chain** | Starts a later stage while prior stage output is missing | Stop and create the missing stage artifact first |

## Self-Correction

From shynlee04/hivemind-plugin, MIT.

More from shynlee04/hivemind-plugin

Signals

Stars
51
Forks
16
Open issues
21
Last commit (this folder)
May 7, 2026 (5 months ago)

From GitHub (fetched Oct 24, 2026).

Details

Version
1.0.0
License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from shynlee04

Done for you

Rather not set it up yourself?

Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.

See how we build AI agents →Read the case studies

Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.

launches shipped
50+
launches shipped
media placements
500+
media placements
founders on every account
2
founders on every account

Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block

Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.

Get hm-l3-research-chain running in your business

Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.

Goes to the Chalk Labs founders, not the newsletter. Privacy

Need help deploying hm-l3-research-chain?

CallGet help
Cite this page

hm-l3-research-chain skill. Agent Mart. https://agentmart.dev/skills/shynlee04-hm-l3-research-chain (last updated 2026-10-24)