CtrlK
BlogDocsLog inGet started
Tessl Logo

sayed/research-pipeline

Produce a rigorous, verified, fully-cited research report on ANY topic by running a disciplined five-stage pipeline: CLARIFY the question, PLAN (an editable research plan), FAN-OUT parallel search subagents, VERIFY every key claim (cross-source + adversarial + cross-model), then SYNTHESIZE a cited report. Use this WHENEVER the user wants real research rather than a quick answer — "research X", "do a deep dive on Y", "write me a report / brief / literature review on Z", "compare A vs B and back it with sources", "what's the state of the art in …", "investigate …", "find evidence for/against …", "give me a market/landscape/competitor analysis", or any question where being wrong is costly and the answer needs citations. Trigger even when the user does not say the word "research" but clearly needs multi-source, fact-checked, cited output. This is the general-topic web-research orchestrator — not the codebase-to-wiki `deep-research` skill. For a fast single-fact lookup, answer directly instead.

75

Quality

94%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Low

Low-risk findings worth noting

Overview
Quality
Evals
Security
Files

Quality

Content

85%Weight 40%Scale 1-3

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The skill body is a well-structured orchestrator playbook: actionable, sequenced with validation, and properly split across real reference files. Its only real weakness is mild verbosity in the motivational framing and per-stage rationales.

Suggestions

Trim the contextual framing ('Every serious research system… is the same loop', the Anthropic token-spend attribution) and condense each stage's '*Why:*' clause to one line to lift conciseness toward a 3.

Move the longer per-stage 'how' detail that currently overlaps with references/playbook.md fully out of SKILL.md so the body stays a lean overview pointing to the references.

DimensionReasoningScore

Conciseness

Most of the body is dense operational guidance Claude would not already know, but it carries contextual padding ('Every serious research system… is the same loop', 'Borrowed from Anthropic's finding…') and verbose per-stage 'why' rationales that could be tightened without losing the actionable signal.

2 / 3

Actionability

Provides copy-paste-ready executable commands (the checkpoint.py invocations), a concrete claim-ledger schema row, a tier table with numeric budgets, and a fan-out decision — all grounded in real referenced scripts (verify_citations.py, checkpoint.py).

3 / 3

Workflow Clarity

A clearly sequenced five-stage pipeline with explicit validation checkpoints (VERIFY, stop-conditions, 'run verify_citations.py before you ship', 'Cap re-search on any single claim at 2 rounds', escalate one tier once) and feedback loops for the fragile verification stage.

3 / 3

Progressive Disclosure

The body is an overview that signals one-level-deep references to real bundle files (references/playbook.md, verification.md, report-template.md, evidence-standards.md and scripts/verify_citations.py, checkpoint.py), each clearly named and scoped in a 'References (read as needed)' section plus inline pointers.

3 / 3

Total

11

/

12

Passed

Description

100%Weight 40%Scale 1-3

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is specific, trigger-rich, complete, and well-differentiated, using imperative third-person voice consistent with the strong reference examples. It is on the longer side but every clause serves a trigger or scoping purpose.

DimensionReasoningScore

Specificity

Names multiple concrete actions — 'Produce a rigorous, verified, fully-cited research report' and an explicit five-stage pipeline (CLARIFY, PLAN, FAN-OUT, VERIFY, SYNTHESIZE) — rather than vague language.

3 / 3

Completeness

Clearly answers both what (produce a verified, cited report via a five-stage pipeline) and when ('Use this WHENEVER the user wants real research rather than a quick answer' plus explicit triggers and a when-not-to-use exclusion).

3 / 3

Trigger Term Quality

Covers a wide range of natural phrases users would say ('research X', 'do a deep dive on Y', 'write me a report / brief / literature review', 'compare A vs B and back it with sources', 'state of the art', 'investigate', 'market/landscape/competitor analysis').

3 / 3

Distinctiveness Conflict Risk

Claims a clear niche as 'the general-topic web-research orchestrator', explicitly distinguishes itself from the codebase-to-wiki 'deep-research' skill, and tells the user to answer directly for fast single-fact lookups, making wrong-skill triggering unlikely.

3 / 3

Total

12

/

12

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation16 / 16 Passed

Validation for skill structure

No warnings or errors.

Reviewed

Table of Contents