Files
accounted/.claude/skills/industry/README.md
T
Jakob Wennberg ec27228a8e style: remove em/en dashes repo-wide, add CLAUDE.md rule against them (#890)
Em dashes (—) and en dashes (–) had spread across comments, docs, tests,
and a few UI strings, reading as AI-generated boilerplate rather than
house style. Replaced each with punctuation matching its context: colon
for explanatory clauses, comma for asides, plain hyphen for numeric/legal
ranges (e.g. "21-23§"), "to"/"till" for date ranges, parentheses for
paired-dash asides. messages/en.json and messages/sv.json were fixed by
hand together to keep sv/en in sync.

Left untouched where the dash is the functional subject rather than
decorative punctuation: date-range-parser.ts's separator regex,
charset-repair.ts's CP1252 byte-mapping table (and its test), the SIE
encoding mojibake docs, generic-csv.ts's minus-sign normalizer, the
agent system-prompt files that already instruct against em dashes, and
a golden iXBRL test fixture compared byte-for-byte.

Also fixes two bugs surfaced along the way: an off-by-one in
ApiKeysPanel's scope-label split (a leftover from an earlier partial
pass), and a charset-repair test that had lost the literal en-dash it
exists to verify.

Regenerated the agent atom seed migration (skills:generate) since 27
SKILL.md files changed. Added a CLAUDE.md rule against em/en dashes,
with an explicit carve-out for the functional-dash cases above.

Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com>
2026-07-04 15:58:06 +02:00

3.0 KiB

Industry (vertical) atoms

Each subdirectory here is one vertical atom: industry-specific knowledge the specialized accountant agent loads when the composer picks it for a company.

.claude/skills/industry/
├── README.md                   ← this file
├── konsult-it/
│   ├── SKILL.md
│   └── references/             ← optional deeper material
│       └── ...
├── restaurang-cafe/
│   └── SKILL.md
└── e-handel/
    └── SKILL.md

Authoring contract

Each SKILL.md must start with YAML frontmatter:

---
id: vertical/<slug>                  # MUST match the directory name
tier: vertical                       # always 'vertical' under industry/
title: "Display name"
description: >
  One paragraph the composer reads when deciding whether to load this atom.
  Cover the activity, the SNI codes it maps to, and the 3-5 most distinctive
  patterns (typical BAS accounts, regulatory edge cases, counterparty types).
sni_prefixes: ["62.01", "62.02", "62.09"]
trigger_signals:
  counterparty_patterns: ["Stripe.*", "AWS.*", "GitHub.*"]
  bas_account_patterns: ["1230", "6230"]
estimated_tokens: 8000               # optional; seeder auto-computes from body
version: 1
---

Body sections (suggested, not enforced):

  1. When to use: short paragraph; helps the agent decide when to lean on this knowledge mid-conversation.
  2. Typical patterns: workflow cadence (invoicing, expenses, reconciliation).
  3. BAS account patterns: the accounts this industry uses heavily.
  4. Regulatory edge cases: quirks specific to the industry (e.g. ROT/RUT for bygg, alkohollagen for restaurang, EU OSS for e-handel).
  5. Counterparties to recognize: common suppliers, payment processors, trade associations.
  6. References: links into deeper references/*.md files in the same dir.

Authoring tips

  • Match the style of the existing horizontal swedish-* skills: developer-oriented, precise, sparse. Not user-facing prose.
  • Quote BAS account numbers as strings ("1465", not 1465).
  • Mention SNI codes explicitly when relevant (SNI 56.10).
  • Avoid duplicating content from the horizontal skills: the composer loads both. If a topic belongs in swedish-vat, link to it ([[horizontal/swedish-vat]]) instead of inlining it.
  • If a section is going to be > 200 lines, split it into a file under references/ and link to it from SKILL.md.

Adding a new atom

  1. Create a new subdirectory under .claude/skills/industry/<slug>/.
  2. Write SKILL.md with the frontmatter above.
  3. Run npx tsx scripts/seed-agent-atom-registry.ts to upsert the atom row into agent_atom_registry.
  4. The composer will start considering the atom on the next signup or "Bygg om" rebuild: no code changes required.

Removing or deprecating

Set is_active: false in frontmatter and re-seed, or set the column directly in the DB. Existing agent_profiles.vertical_atoms arrays may still reference the inactive ID; the runtime loader skips inactive atoms.