Parse a single worksheet into a list of company entries and detected categories.

Compute a match score between 0 and 100 for ranking results.

Split a command spec into {heading: body} by '##' headers. Splitting this way lets a fork's extra sections (e.g. this fork's '## Blocker logg

Search for a company by name. Returns matching entries sorted by relevance.

Body of one markdown section, up to the next heading of any depth.

Return a human-readable problem per structural defect in [Unreleased].

Split a skill spec into {heading: body} by '##' headers. Splitting this way lets a fork's extra sections sit between the ones under test with

Format a single company entry for display.

The body of one markdown section, up to the next heading of any depth.

Detect whether a column header refers to count or index data.

Does this actually look like a robots.txt? A misconfigured host can answer /robots.txt with 200 and an HTML error page. That body parses to z

Normalize string for robust fuzzy matching.

Parse numeric Excel values, including localized string cells.

Convert Danish/Nordic characters to anglicized equivalents.

Return (errors, warnings) for the salary data shape. errors -> hard problems that make lookups crash or emit wrong output (thes

Extract meaningful words from a company name, ignoring noise.

The slice of text from the start marker up to the end marker.

apply.md's '### 1. Research the Company' subsection, isolated from the other numbered subsections under Step 3.

Split a markdown spec into {heading: body} on a given '\\n<marker> ' prefix.

Extract ATS-readable text. Returns (text, pages, extractor_name).

Return True when a header contains a meaningful pattern match. Patterns match whole tokens; any pattern also listed in ``COMPOUND_PATTERNS``

Return (document, seen-map). The map is mutated in place by callers.

Validate the salary data shape before lookups use it. Preserves historical behavior: exits(1) on the first hard error with the same user-faci

Run main() against `payload` with the given argv. Returns (exit_code_or_None, stdout). main() returns normally on a successful render,

Split a command spec into {heading: body} by '## ' headers.

Exit with a user-facing salary data setup error.

Compute a match score between 0 and 100 using precalculated query values.

Rule 6's defensive-parse rule: anything that is not YYYY-MM-DD is treated exactly like an absent value - never compared, never guessed at.

Load and JSON-parse salary_data.json; exit with a helpful message if missing/invalid.

Atomic replace: a half-written seen_jobs.json loses the scrape history.

Return the body of a markdown section up to the next heading.

Body of a markdown section up to the next heading of the same level.

Return (text, pages) or None if pypdf is unavailable, raises, or yields no text.

curl, not urllib: some hosts (jobup.ch) hang urllib indefinitely while answering curl in under a second, and --max-time is a hard ceiling.

user-agent -> [(is_allow, pattern)], tolerating blank lines inside a record.

Yield "<Event>:<command>" for every command a hook event would run. Fails closed: any shape this does not recognise yields a marker that cannot

RFC 9309 wildcard match; returns match length or -1. The pattern is percent-decoded to match the already-decoded path. Without this, "Disallo

The field names Step 2's scoring-agent JSON contract declares. Extracted from the fenced ```json block in Step 2 rather than hardcode

Split the skill spec into {heading: body} by '###' step headers. Splitting this way lets a fork's extra steps sit between the ones under test

Split a step body into {'### N. ...' heading: body}.

The Subfolder naming rule, executed exactly as documented: lowercase, underscores for spaces, drop every character that is not a lette

location_verdict, falling back to a legacy verdict stored under `location` (Step 4: "an entry ranked before this rename may carry a legacy PASS/FA