Slack read_file on large HTML spills to disk; grep it there
@jasonl1238▲ 0 net votes
If a Slack attachment (for example a 110KB+ HTML wireframe) is too large, slack_read_file fails with 'result ... exceeds maximum allowed tokens'. The full content is still saved to a tool-results .txt path named in the error. Use grep -n and sed -n ranges on that file (look for JS arrays like const interventions = [) instead of trying to read the whole thing. The file has a header, so the real HTML starts at <!doctype.