research

Name: research
Author: JuliusBrussee

450installs81forks

How to install research?

juliusbrussee/cavekit/research

$npx -y skills add JuliusBrussee/cavekit --skill research

Installs into the current project.

Summary

Gather external knowledge the spec needs and distill it into §R — the durable research log — so build grounds in facts instead of hallucinating library behavior. Each finding cites a source; unsourced claims are flagged, never written as fact. Triggers when a spec decision hinges

Files · 1

View on GitHub

SKILL.md

1	# research — external knowledge → §R
2
3	Every finding cites a source. No source → flag it `?`, never write a guess as fact.
4
5	"Process without library context gives you well-organized hallucinations."
6	Build invents a plausible-but-wrong API & §B fills with avoidable bugs.
7	Research is the external oracle: pull the real fact once, log it caveman, never re-derive.
8
9	## WHEN TO RESEARCH
10
11	- A §C/§I/§V decision hinges on a lib, API, version, or pattern you are unsure of.
12	- You are about to assume how an external dependency behaves.
13	- The idea touches a domain with real prior art (auth, payments, crypto, rate-limit).
14	- `/grill` parked a `?` that the outside world must answer.
15
16	Skip when the build touches only code you already wrote. Research scales to the unknown, ⊥ to habit.
17
18	## FOUR STEPS
19
20	### 1. SCOPE
21	Turn the unknown into 1-3 concrete questions. Vague "research auth" → "JWT lib
22	for Node ESM, maintained?" + "refresh-token rotation: current best practice?".
23	A scoped question gets a citable answer; a vague one gets an essay.
24
25	### 2. GATHER
26	Use web search / docs tools. Prefer primary sources: official docs, the repo,
27	the RFC, the paper. Two independent sources beat one confident blog. For a big
28	sweep, spawn a sub-agent so the raw pages never touch this context — it returns
29	only the distilled finding + source.
30
31	### 3. DISTILL
32	Crush each answer to one caveman line + its source. Drop the prose. The §R row
33	is the memory; the tab you read is not.
34
35	> R3\|refresh token\|rotate on use, revoke family on reuse-detect\|datatracker.ietf.org/doc/html/rfc6819#section-5.2.2.3
36
37	### 4. HAND OFF
38	Emit the §R rows & hand to the spec skill to append. If a finding changes a
39	constraint or interface, note the §C/§I edit for spec too. Research proposes;
40	spec writes.
41
42	## SOURCE DISCIPLINE
43
44	- Cite a URL, repo, RFC, or paper per row. Verbatim identifiers/versions.
45	- Could not verify → write the row but flag `?` in the finding & say so. An
46	unverified claim labeled honestly is fine; one disguised as fact is a future §B.
47	- Conflicting sources → log both, let the user pick. ⊥ silently average them.
48
49	## WHEN TO STOP
50
51	Done when every scoped question has a sourced §R row (or an honest `?`), and no
52	build decision still rests on an unchecked assumption. ⊥ research past the
53	questions you scoped — that is just burning the attention budget.
54
55	## BOUNDARIES
56
57	- ⊥ write SPEC.md. Hand §R rows to spec.
58	- ⊥ write a finding as fact without a source.
59	- ⊥ dump raw pages into context or §R. Distill or it does not land.
60	- ⊥ research what you can read in the repo. Local truth > web guess.

People who liked this also viewed

6 picks

Type

research

450installs81forks

View on GitHub

How to install research?

juliusbrussee/cavekit/research

$npx -y skills add JuliusBrussee/cavekit --skill research

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# research — external knowledge → §R
2
3	Every finding cites a source. No source → flag it `?`, never write a guess as fact.
4
5	"Process without library context gives you well-organized hallucinations."
6	Build invents a plausible-but-wrong API & §B fills with avoidable bugs.
7	Research is the external oracle: pull the real fact once, log it caveman, never re-derive.
8
9	## WHEN TO RESEARCH
10
11	- A §C/§I/§V decision hinges on a lib, API, version, or pattern you are unsure of.
12	- You are about to assume how an external dependency behaves.
13	- The idea touches a domain with real prior art (auth, payments, crypto, rate-limit).
14	- `/grill` parked a `?` that the outside world must answer.
15
16	Skip when the build touches only code you already wrote. Research scales to the unknown, ⊥ to habit.
17
18	## FOUR STEPS
19
20	### 1. SCOPE
21	Turn the unknown into 1-3 concrete questions. Vague "research auth" → "JWT lib
22	for Node ESM, maintained?" + "refresh-token rotation: current best practice?".
23	A scoped question gets a citable answer; a vague one gets an essay.
24
25	### 2. GATHER
26	Use web search / docs tools. Prefer primary sources: official docs, the repo,
27	the RFC, the paper. Two independent sources beat one confident blog. For a big
28	sweep, spawn a sub-agent so the raw pages never touch this context — it returns
29	only the distilled finding + source.
30
31	### 3. DISTILL
32	Crush each answer to one caveman line + its source. Drop the prose. The §R row
33	is the memory; the tab you read is not.
34
35	> R3\|refresh token\|rotate on use, revoke family on reuse-detect\|datatracker.ietf.org/doc/html/rfc6819#section-5.2.2.3
36
37	### 4. HAND OFF
38	Emit the §R rows & hand to the spec skill to append. If a finding changes a
39	constraint or interface, note the §C/§I edit for spec too. Research proposes;
40	spec writes.
41
42	## SOURCE DISCIPLINE
43
44	- Cite a URL, repo, RFC, or paper per row. Verbatim identifiers/versions.
45	- Could not verify → write the row but flag `?` in the finding & say so. An
46	unverified claim labeled honestly is fine; one disguised as fact is a future §B.
47	- Conflicting sources → log both, let the user pick. ⊥ silently average them.
48
49	## WHEN TO STOP
50
51	Done when every scoped question has a sourced §R row (or an honest `?`), and no
52	build decision still rests on an unchecked assumption. ⊥ research past the
53	questions you scoped — that is just burning the attention budget.
54
55	## BOUNDARIES
56
57	- ⊥ write SPEC.md. Hand §R rows to spec.
58	- ⊥ write a finding as fact without a source.
59	- ⊥ dump raw pages into context or §R. Distill or it does not land.
60	- ⊥ research what you can read in the repo. Local truth > web guess.

People who liked this also viewed

6 picks

Type