slacrawl

# Slacrawl Use local Slack archive data first. Hit Slack APIs only when archive is stale, missing scope, or user asks for current external context. ## Sources - DB: `~/.slacrawl/slacrawl.db` - Config: `~/.slacrawl/config.toml` - Cache/logs: `~/.slacrawl/{cache,logs}` - Git share repo: `~/.slacrawl/share` - Repo: `~/Projects/slacrawl` - CLI: `slacrawl` ## Freshness For recent/current Slack questions: ```bash slacrawl doctor slacrawl status --json ``` Refresh: ```bash slacrawl sync --source bot --latest-only slacrawl sync --source mcp --workspace T01234567 slacrawl sync --source wiretap ``` Use `--full` only for deliberate historical backfills. `bot` = API tokens; `mcp` = configured HTTP or stdio Slack connector; `wiretap` = Slack Desktop cache; `all` = API then desktop enrichment. ## Query Workflow 1. Resolve workspace, channel/DM, date range, user, and keyword. 2. Check freshness if the question is recent/current. 3. Prefer CLI search/messages for slices; use read-only SQL for exact counts. 4. Report workspace/channel names, date spans, counts, and token/source limits. Use root or subcommand help for syntax: `slacrawl --help`, `slacrawl search --help`, `slacrawl messages --help`, `slacrawl sql --help`. `slacrawl search --limit N` is supported. Common commands: ```bash slacrawl search --limit 20 "query" slacrawl messages --since 7d --limit 50 slacrawl channels --json slacrawl users --json slacrawl mentions --limit 50 slacrawl sql 'select count(*) from messages;' ``` Use `slacrawl --json sql ...` for exact read-only counts/joins/rankings. Keep SQL to `select`/`with`. ## Verification For repo edits: ```bash GOWORK=off go test ./... make test ``` Use a small CLI smoke such as: ```bash slacrawl doctor slacrawl search test ```