weekly-digest

/weekly — Weekly Industry Digest Skill (v1.0)

Produces a SEO- and GEO-optimised weekly news digest article, published as DRAFT by default (Rule 5a). The pipeline reuses the standard run_pipeline loop from skills/seo-blog/SKILL.md for all build→publish stages, but bypasses keyword-research, SERP-analysis, and competitor-analysis (meaningless for a digest) via explicit --action skip.

⛔ READ FIRST — Rule 5a: DRAFT by default

Every /weekly run creates a WordPress post with status: "draft".
Publishing live requires explicit user confirmation in the same conversation: "publish" / "go live" / "发布" / "上线" — or a project-level publish_policy.default = "publish" in business-context.json.
The pipeline MUST NOT flip status to publish without that signal.

Step 0 — Startup: resolve active project (env-first)

active_slug=$(python -c "from scripts._core import active_project as a; print(a.get_active_project() or '')")

If active_slug is empty, ask the user which project to run for (/switch <slug> or set XS_ACTIVE_PROJECT env var). Do not proceed without a slug.

Step 1 — Load weekly_digest config; stop if absent or disabled

python -c "
import json, sys
from pathlib import Path
bc = json.loads(Path('projects/{slug}/business-context.json').read_text(encoding='utf-8'))
wd = bc.get('weekly_digest', {})
if not wd.get('enabled'):
    print('NOT_ENABLED')
    sys.exit(2)
print('OK')
"

If the command prints NOT_ENABLED (or the file is missing), stop and tell the user:

The weekly_digest block is absent or enabled: false in projects/{slug}/business-context.json. Add a weekly_digest object with at least {"enabled": true, "series_keyword": "...", "connectors": {...}} to activate the feature. Plan 3 (loamwright rollout) wires the first production config.

Step 2 — Create task workspace

tid = "{slug}_wk_{YYYYMMDD}"   # e.g. loamwright_wk_20260630
                                # must match ^[a-z0-9_]{8,32}$
ws  = "memory/workspace/{tid}/"