monte-carlo-data/mc-agent-toolkit / .claude/skills/toolkit-skill-author
toolkit-skill-author
VerifiedAuthors or extends a skill in mc-agent-toolkit. Gates for forbidden buckets and name collisions, applies CONTRIBUTING's extend-or-split rules, then edits a peer skill or hands off to Anthropic's skill-creator and walks the registration checklist.
| File | Kind | Size |
|---|---|---|
| README.md | markdown | 5.7 KB |
| SKILL.md | skill md | 8.9 KB |
| references/decision-rules.md | markdown | 3.2 KB |
| references/handoff-preamble.md | markdown | 3.8 KB |
| references/registration-checklist.md | markdown | 4.7 KB |
| scripts/find-peers.shexecutable | script | 1020 B |
| scripts/lint-skill.pyexecutable | script | 8.0 KB |
| scripts/tests/find-peers.bats | other | 1.2 KB |
| scripts/tests/lint-skill.bats | other | 7.9 KB |
| scripts/tests/smoke-test.shexecutable | script | 2.4 KB |
| scripts/tests/test_portable_frontmatter.py | script | 1.8 KB |
scripts/tests/lint-skill.bats
View on GitHub#!/usr/bin/env bats
setup() {
SCRIPT="$BATS_TEST_DIRNAME/../lint-skill.py"
TMP=$(mktemp -d)
}
teardown() {
rm -rf "$TMP"
}
# Write a SKILL.md with given frontmatter under $TMP/<dir>/SKILL.md
make_skill() {
local dir="$1"
local frontmatter="$2"
mkdir -p "$TMP/$dir"
cat > "$TMP/$dir/SKILL.md" <<EOF
---
$frontmatter
---
Body content.
EOF
}
@test "passes for a fully-compliant skill" {
make_skill good-skill "name: monte-carlo-good-skill
description: Does a thing for users when they ask about doing that thing.
when_to_use: |
When the user mentions doing that thing.
bucket: Monitoring"
run python3 "$SCRIPT" good-skill "$TMP"
[ "$status" -eq 0 ]
[[ "$output" == *"good-skill"* ]]
[[ "$output" != *"ERROR"* ]]
}
@test "passes for a block-scalar description" {
make_skill block-skill 'name: monte-carlo-block-skill
description: |
This is a multi-line description that spans
several lines for readability.
when_to_use: |
When the user needs this.
bucket: Trust'
run python3 "$SCRIPT" block-skill "$TMP"
[ "$status" -eq 0 ]
[[ "$output" != *"ERROR"* ]]
}
@test "errors when description is missing" {
make_skill nodesc "name: monte-carlo-nodesc
when_to_use: Something.
bucket: Monitoring"
run python3 "$SCRIPT" nodesc "$TMP"
[ "$status" -eq 1 ]
[[ "$output" == *"ERROR: description is missing"* ]]
}
@test "errors when description exceeds 1024 chars" {
long=$(python3 -c 'print("x" * 1100)')
make_skill toolong "name: monte-carlo-toolong
description: $long
bucket: Monitoring"
run python3 "$SCRIPT" toolong "$TMP"
[ "$status" -eq 1 ]
[[ "$output" == *"description is 1100 chars"* ]]
}
@test "errors on first-person opener 'This skill'" {
make_skill firstperson "name: monte-carlo-firstperson
description: This skill does a thing when asked.
bucket: Monitoring"
run python3 "$SCRIPT" firstperson "$TMP"
[ "$status" -eq 1 ]
[[ "$output" == *"first-person"* ]]
}
@test "errors on first-person opener 'Use this skill'" {
make_skill useskill "name: monte-carlo-useskill
description: Use this skill when the user wants X.
bucket: Monitoring"
run python3 "$SCRIPT" useskill "$TMP"
[ "$status" -eq 1 ]
[[ "$output" == *"first-person"* ]]
}
@test "warns when when_to_use is missing, still exits 0" {
make_skill nowtu "name: monte-carlo-nowtu
description: A valid description.
bucket: Monitoring"
run python3 "$SCRIPT" nowtu "$TMP"
[ "$status" -eq 0 ]
[[ "$output" == *"WARN: when_to_use is missing"* ]]
}
@test "errors when combined description + when_to_use exceeds 1400 chars" {
desc=$(python3 -c 'print("a" * 900)')
wtu=$(python3 -c 'print("b" * 600)')
make_skill combined "name: monte-carlo-combined
description: $desc
when_to_use: $wtu
bucket: Monitoring"
run python3 "$SCRIPT" combined "$TMP"
[ "$status" -eq 1 ]
[[ "$output" == *"1500 chars"* ]]
[[ "$output" == *"max 1400"* ]]
}
@test "errors when name is bare directory without monte-carlo- prefix" {
make_skill bare-name "name: bare-name
description: Does a thing.
bucket: Monitoring"
run python3 "$SCRIPT" bare-name "$TMP"
[ "$status" -eq 1 ]
[[ "$output" == *"should be 'monte-carlo-bare-name'"* ]]
}
@test "errors when name field has wrong prefix" {
make_skill my-dir "name: other-prefix-my-dir
description: Does a thing.
bucket: Monitoring"
run python3 "$SCRIPT" my-dir "$TMP"
[ "$status" -eq 1 ]
[[ "$output" == *"should be 'monte-carlo-my-dir'"* ]]
}
@test "errors when name is not kebab-case" {
make_skill bad-dir "name: BadName
description: Does a thing.
bucket: Monitoring"
run python3 "$SCRIPT" bad-dir "$TMP"
[ "$status" -eq 1 ]
[[ "$output" == *"not kebab-case"* ]]
}
@test "errors when name contains underscores" {
make_skill bad-dir2 "name: monte_carlo_bad_dir2
description: Does a thing.
bucket: Monitoring"
run python3 "$SCRIPT" bad-dir2 "$TMP"
[ "$status" -eq 1 ]
[[ "$output" == *"not kebab-case"* ]]
}
@test "errors when bucket is missing" {
make_skill nobucket "name: monte-carlo-nobucket
description: Does a thing.
when_to_use: When needed."
run python3 "$SCRIPT" nobucket "$TMP"
[ "$status" -eq 1 ]
[[ "$output" == *"bucket is missing"* ]]
}
@test "errors when bucket is invalid" {
make_skill badbucket "name: monte-carlo-badbucket
description: Does a thing.
when_to_use: When needed.
bucket: Nonsense"
run python3 "$SCRIPT" badbucket "$TMP"
[ "$status" -eq 1 ]
[[ "$output" == *"bucket 'Nonsense' is not a valid capability bucket"* ]]
}
@test "accepts multi-word bucket 'Incident Response'" {
make_skill ir-skill "name: monte-carlo-ir-skill
description: Does a thing.
when_to_use: When needed.
bucket: Incident Response"
run python3 "$SCRIPT" ir-skill "$TMP"
[ "$status" -eq 0 ]
[[ "$output" != *"ERROR"* ]]
}
@test "warns when version field is present, still exits 0" {
make_skill withversion "name: monte-carlo-withversion
description: Does a thing.
when_to_use: When needed.
bucket: Monitoring
version: 1.0.0"
run python3 "$SCRIPT" withversion "$TMP"
[ "$status" -eq 0 ]
[[ "$output" == *"WARN:"* ]]
[[ "$output" == *"version field in SKILL.md is ignored"* ]]
}
@test "exits 2 when SKILL.md does not exist" {
run python3 "$SCRIPT" nonexistent "$TMP"
[ "$status" -eq 2 ]
[[ "$output" == *"not found"* ]]
}
@test "exits 2 with usage message when no args given" {
run python3 "$SCRIPT"
[ "$status" -eq 2 ]
[[ "$output" == *"Usage"* ]]
}
@test "handles quoted scalar descriptions" {
make_skill quoted 'name: monte-carlo-quoted
description: "A quoted description with colons: and commas."
bucket: Monitoring'
run python3 "$SCRIPT" quoted "$TMP"
[ "$status" -eq 0 ]
[[ "$output" == *"quoted"* ]]
[[ "$output" != *"ERROR"* ]]
}
@test "defaults skills_root to ./skills when omitted" {
mkdir -p "$TMP/skills/defroot"
cat > "$TMP/skills/defroot/SKILL.md" <<'EOF'
---
name: monte-carlo-defroot
description: A default-root test skill.
bucket: Monitoring
---
EOF
cd "$TMP"
run python3 "$SCRIPT" defroot
[ "$status" -eq 0 ]
[[ "$output" == *"defroot"* ]]
}
@test "errors on unclosed plugin-only start marker" {
mkdir -p "$TMP/unclosed"
cat > "$TMP/unclosed/SKILL.md" <<'EOF'
---
name: monte-carlo-unclosed
description: Does a thing.
bucket: Monitoring
---
Body content.
<!-- plugin-only:start -->
Plugin-only content that is never closed.
EOF
run python3 "$SCRIPT" unclosed "$TMP"
[ "$status" -eq 1 ]
[[ "$output" == *"plugin-only:start never closed"* ]]
}
@test "errors on orphan plugin-only end marker" {
mkdir -p "$TMP/orphan-end"
cat > "$TMP/orphan-end/SKILL.md" <<'EOF'
---
name: monte-carlo-orphan-end
description: Does a thing.
bucket: Monitoring
---
Body content.
<!-- plugin-only:end -->
EOF
run python3 "$SCRIPT" orphan-end "$TMP"
[ "$status" -eq 1 ]
[[ "$output" == *"plugin-only:end without an open"* ]]
}
@test "errors on nested plugin-only start marker" {
mkdir -p "$TMP/nested"
cat > "$TMP/nested/SKILL.md" <<'EOF'
---
name: monte-carlo-nested
description: Does a thing.
bucket: Monitoring
---
<!-- plugin-only:start -->
Outer block.
<!-- plugin-only:start -->
Inner block.
<!-- plugin-only:end -->
EOF
run python3 "$SCRIPT" nested "$TMP"
[ "$status" -eq 1 ]
[[ "$output" == *"nested plugin-only:start"* ]]
}
@test "errors when a plugin-only marker shares a line with other text" {
mkdir -p "$TMP/shared"
cat > "$TMP/shared/SKILL.md" <<'EOF'
---
name: monte-carlo-shared
description: Does a thing.
bucket: Monitoring
---
Body content. <!-- plugin-only:start --> trailing text.
EOF
run python3 "$SCRIPT" shared "$TMP"
[ "$status" -eq 1 ]
[[ "$output" == *"not line-anchored"* ]]
}
@test "passes with a correctly paired plugin-only block" {
mkdir -p "$TMP/paired"
cat > "$TMP/paired/SKILL.md" <<'EOF'
---
name: monte-carlo-paired
description: Does a thing.
bucket: Monitoring
---
Body content.
<!-- plugin-only:start -->
/run-a-command and read references/foo.md
<!-- plugin-only:end -->
Portable content again.
EOF
run python3 "$SCRIPT" paired "$TMP"
[ "$status" -eq 0 ]
[[ "$output" != *"ERROR"* ]]
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 470k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |