fp-go Pattern Matching

What it does

When your agent uses it

• Replacing a Go switch or if-else chain with composable pattern matching
• Writing a router, classifier, or multi-format parser in fp-go v2
• Reviewing fp-go code that chains Alt, FindFirstMap, or similar combinators

Example prompts

• “Rewrite this switch statement as a fp-go point-free match.”
• “Add a fallback case for an unrecognized request format.”
• “Review this classifier for point-free style and argument-order mistakes.”

Requirements

• Go with the github.com/IBM/fp-go/v2 package

Workflow steps

8 steps, taken from the first numbered list in SKILL.md.

1. 1Cases are ordered specific -> general (a general case first shadows later ones).
2. 2No AltAllArray([]Option{c(x), ...}) over case applications. Rewrite to A.Fold(RO.AltMonoid...).
3. 3Point-free: no lambdas that only compare a field, call one stdlib function,
4. 4The matcher is built by a constructor func, once, not inside a loop or per request.
5. 5Intended total functions end in RO.GetOrElse, so Option does not leak.
6. 6Cases are pure. If a branch needs IO, return the effect (Option[IOResult[B]]) instead of running it in the guard.
7. 7Each case has a unit test, plus a test of the combined match, including the default and priority.
8. 8Fixed, local, non-reused branching stays a plain switch. Do not force this pattern where it adds nothing.

What it can do on your machine

Read from SKILL.md and the folder at commit 1c4245d. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  Shell commands in SKILL.md call:

  • go

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  No URLs in SKILL.md.

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

fp-go Pattern Matching loads about 2.8k tokens when it runs. Until then it costs about 168 tokens; SKILL.md has 760 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~168

When it runs · the whole SKILL.md, loaded when a task matches

~2.8k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from IBM/fp-go at commit 1c4245d, republished under its Apache-2.0 licence (© IBM). 760 words, ~2,782 tokens.

name

fp-go-pattern-matching

description

Use this skill when writing, refactoring, or reviewing fp-go v2 code that replaces switch/case, if-else chains, or type switches with composable, point-free pattern matching: first-match-wins case lists, routers, classifiers, multi-format parsers, fallback chains, or matching over sum types and slices. Trigger on any mention of pattern matching in Go, match/case, switch to functional, AltAllArray, AltAllSeq, AltMonoid, readeroption.AltMonoid, readeroption.Alt, FindFirstMap, FindLastMap, FilterMap, option.Alt, first Some, fallback chain, InstanceOf, prism GetOption as matcher, GetOrElse default branch, option.Fold, either.Fold, or array.Match.

import (
    A  "github.com/IBM/fp-go/v2/array"
    F  "github.com/IBM/fp-go/v2/function"
    N  "github.com/IBM/fp-go/v2/number"
    O  "github.com/IBM/fp-go/v2/option"
    P  "github.com/IBM/fp-go/v2/predicate"
    R  "github.com/IBM/fp-go/v2/result"
    RD "github.com/IBM/fp-go/v2/reader"
    RO "github.com/IBM/fp-go/v2/readeroption"
    S  "github.com/IBM/fp-go/v2/string"
    "github.com/IBM/fp-go/v2/optics/prism"
    "github.com/IBM/fp-go/v2/readerresult" // generic ReaderResult[T, B]; not context/readerresult
)

func getMethod(r Request) string { return r.Method }   // leaf accessor (or lens.Get)
func getPath(r Request) string   { return r.Path }

func on(guard P.Predicate[Request], branch func(Request) string) RO.ReaderOption[Request, string] {
    return F.Flow2(O.FromPredicate(guard), O.Map(branch))
}

func isMethod(m string) P.Predicate[Request] {
    return F.Flow2(getMethod, P.IsStrictEqual[string]()(m))
}

func HandleRequest() func(Request) string {
    return F.Pipe2(
        A.From(                                           // ordered, specific -> general
            on(isMethod("GET"), F.Flow2(getPath, S.Format[string]("Fetching: %s"))),
            on(isMethod("DELETE"), F.Flow2(getPath, S.Format[string]("Deleting: %s"))),
        ),
        A.Fold(RO.AltMonoid[Request, string]()),          // lazy first-match
        RO.GetOrElse(F.Flow2(getMethod, S.Format[string]("Unsupported: %s"))), // default
    )
}

func when[T, B any](guard P.Predicate[T], b B) RO.ReaderOption[T, B] {
    return F.Flow2(O.FromPredicate(guard), O.Map(F.Constant1[T](b)))
}

// append one or two fallback cases
F.Pipe2(c1, RO.Alt(F.Constant(c2)), RO.Alt(F.Constant(c3)))

// FindFirstMap over a slice of cases
F.Flow3(
    RD.Read[O.Option[B], T],                   // x -> apply-a-case-to-x
    A.FindFirstMap[RO.ReaderOption[T, B], B], // -> first case that matches
    RD.Read[O.Option[B]](cases),               // run over the cases
)

// BAD: all three cases run, because Go evaluates the slice literal first
O.AltAllArray(O.None[B]())([]O.Option[B]{c1(x), c2(x), c3(x)})
// its point-free equivalent shows why: SequenceArray runs every case first
F.Pipe2(cases, RD.SequenceArray[T, O.Option[B]], RD.Map[T](O.AltAllArray(O.None[B]())))

func withPrefix(prefix string, base int) RO.ReaderOption[string, int64] {
    return F.Flow3(
        O.FromPredicate(F.Bind2nd(strings.HasPrefix, prefix)),
        O.Map(F.Bind2nd(strings.TrimPrefix, prefix)),
        O.Chain(F.Flow2(F.Bind23of3(R.Eitherize3(strconv.ParseInt))(base, 64), R.ToOption[int64])),
    )
}

func caseOf[T Shape, B any](f func(T) B) RO.ReaderOption[Shape, B] {
    return F.Flow3(F.ToAny[Shape], O.InstanceOf[T], O.Map(f))
}
describe := A.Fold(RO.AltMonoid[Shape, string]())(A.From(
    caseOf(S.Format[Circle]("circle %v")),
    caseOf(S.Format[Rect]("rect %v")),
))

F.Pipe1(match, RO.GetOrElse(F.Flow2(getX, S.Format[X]("no match: %v")))) // default sees input
F.Pipe1(match, RO.GetOrElse(F.Constant1[T](dflt)))                       // constant default
A.From(c1, c2, RO.Of[T](dflt))                                           // catch-all case, still Option

func withPrefixR(prefix string, base int, err error) readerresult.ReaderResult[string, int64] {
    return F.Flow4(
        R.FromPredicate(F.Bind2nd(strings.HasPrefix, prefix), F.Constant1[string](err)),
        R.Map(F.Bind2nd(strings.TrimPrefix, prefix)),
        R.Chain(F.Bind23of3(R.Eitherize3(strconv.ParseInt))(base, 64)),
        R.MapLeft[int64](F.Constant1[error](err)),
    )
}

parse := A.Fold(readerresult.AltMonoid(F.Constant(readerresult.Left[string, int64](errNoMatch))))(A.From(
    withPrefixR("0x", 16, errHex),
    withPrefixR("", 10, errDec),
))
// two alternatives: F.Pipe1(c1, readerresult.Alt(F.Constant(c2)))

© IBM, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in skills/fp-go-pattern-matching of IBM/fp-go.

Open the folder on GitHubat commit 1c4245d