@m13v/s4l 1.7.6-rc.34 → 1.7.6-rc.35

package/mcp/dist/version.json

@@ -1,4 +1,4 @@
 {
-  "version": "1.7.6-rc.34",
-  "installedAt": "2026-07-18T02:38:17.233Z"
+  "version": "1.7.6-rc.35",
+  "installedAt": "2026-07-18T02:51:25.957Z"
 }

package/mcp/manifest.json

@@ -2,7 +2,7 @@
   "dxt_version": "0.1",
   "name": "social-autoposter",
   "display_name": "S4L",
-  "version": "1.7.6-rc.34",
+  "version": "1.7.6-rc.35",
   "description": "Draft, review, approve, and autopilot X/Twitter posts.",
   "long_description": "## **⚠️ The disclaimer above is generic Claude boilerplate.** Anthropic shows the same warning on every plugin regardless of what it does; any plugin has the same level of access as any app you download from the internet.\n\nS4L is an open source product developed by Mediar.ai Incorporated, a VC-backed San Francisco-based startup.\n\nTo get started:\n\n1\\. Copy this prompt: **Set me up on S4L plugin end to end**\n\n2\\. Quit with CMD+Q, reopen Claude, paste into a new chat.\n\nWhat happens next:\n\n* About every 5 minutes S4L scans X for posts that match your topics and drafts replies in your voice.\n* Drafts show up as review cards, usually the first within a few minutes. Nothing is posted automatically; you approve each one.\n* Posting autopilot stays off until you explicitly turn it on.",
   "author": {

package/mcp/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@m13v/s4l-mcp",
-  "version": "1.7.6-rc.34",
+  "version": "1.7.6-rc.35",
   "private": true,
   "description": "Desktop MCP client for social-autoposter (X/Twitter rail): manual draft/review/approve loop, autopilot control, and stats. Thin wrapper over the existing pipeline scripts.",
   "license": "MIT",

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@m13v/s4l",
-  "version": "1.7.6-rc.34",
+  "version": "1.7.6-rc.35",
   "description": "Automated social posting pipeline for Reddit, X/Twitter, LinkedIn, and Moltbook. Install as a Claude Code agent skill.",
   "bin": {
     "social-autoposter": "bin/cli.js",

package/scripts/draft_prompt_core.py

@@ -688,10 +688,14 @@ def render_reddit_prompt(ing):
     recent_self_block, arm (optional), lane (optional, promotion assumed).
     """
     arm = ing.get("arm") or os.environ.get("S4L_DRAFT_PROMPT_VARIANT") or ""
-    # Reddit drafting is promotion-lane work by definition today (one product
-    # project per draft call); the persona directive would reference blocks
-    # this prompt does not carry. Force the promotion directive regardless of
-    # any stray S4L_ACTIVE_LANE in the environment.
+    # Lane comes ONLY from the ingredients (post_reddit.py stamps it onto the
+    # discover plan at the cycle's assignment point and threads it through
+    # build_draft_prompt). Deliberately NOT read from S4L_ACTIVE_LANE here:
+    # salvage plans drafted in a later cycle must keep their own (promotion)
+    # directive even when THIS cycle's coin flip landed personal_brand.
+    # lane='personal_brand' renders the persona directive + the persona-
+    # whitelisted projects_json; the RD template carries every block the
+    # persona directive references (@CORPUS_BLOCK@ = ACCOUNT VOICE CORPUS).
     lane = ing.get("lane") or ""
 
     from engagement_styles import get_content_rules, get_voice_relationship_rule

package/scripts/link_tail.py

@@ -557,7 +557,7 @@ def main() -> int:
         print(json.dumps(out), flush=True)
         return 0
 
-    # Plugin (MCP post_drafts) flow sets S4L_SKIP_LINK_TAIL=1. The bridge only
+    # Plugin (MCP approve_drafts) flow sets S4L_SKIP_LINK_TAIL=1. The bridge only
     # rewords prose around the URL — the minted short link is produced by a
     # separate deterministic wrap step in twitter_post_plan.py — so the Claude
     # call buys nothing there, and on .mcpb customer boxes (no `claude` binary)

package/scripts/merge_review_queue.py

@@ -235,7 +235,7 @@ def _sync_with_backend(cands: list) -> tuple[int, int]:
         skill/run-twitter-cycle.sh) and mark it terminal here too, same as a
         human "discard all pending" would. Without this, a card can sit in
         the review queue as an approvable draft long after the backend has
-        moved on; approving it later silently no-ops (post_drafts returns
+        moved on; approving it later silently no-ops (approve_drafts returns
         posted:0, no browser ever launches, no post-*.log — see the
         2026-07-09 "approved 3 cards, nothing posted" investigation).
 
@@ -246,7 +246,7 @@ def _sync_with_backend(cands: list) -> tuple[int, int]:
 
     APPROVED cards are exempt from the prune: an approval is a settled human
     decision awaiting its serialized post, and stamping terminal on it makes
-    post_drafts refuse it as already-decided (posted:0). That exact race ate
+    approve_drafts refuse it as already-decided (posted:0). That exact race ate
     2 of 3 approvals on 2026-07-10 (approved 04:30:02Z, merge stamped
     backend_status_expired 04:32:33Z, poster refused both). The freshness
     gate exists to stop stale UNDECIDED drafts from burning review attention;

package/scripts/post_reddit.py

@@ -442,8 +442,21 @@ def _db_pick_salvage_candidates(batch_id, limit=1):
         data = (resp or {}).get("data") or {}
         if not data.get("decisions"):
             return None
+        # Lane for salvage rows is derived from CONFIG (persona:true on the
+        # row's project), not from this cycle's env: a persona row that
+        # failed transiently must re-draft under the persona directive even
+        # when the retry cycle's coin flip landed promotion, and vice versa.
+        _sv_proj = (data.get("project_name") or "").lower()
+        try:
+            _sv_lane = "personal_brand" if any(
+                p.get("persona") is True and (p.get("name") or "").lower() == _sv_proj
+                for p in (load_config().get("projects") or [])
+            ) else "promotion"
+        except Exception:
+            _sv_lane = "promotion"
         return {
             "project_name": data.get("project_name") or "general",
+            "lane": _sv_lane,
             # batch_id must ride the plan file into --phase draft: the
             # snapshot filename, the prompt's top_performers --invoked-by,
             # and the proposed_excludes 2-distinct-batch activation gate all
@@ -1204,7 +1217,7 @@ def _prefetch_thread_digests(candidates, reddit_username=None):
 
 def build_draft_prompt(project, config, candidates, top_report, recent_comments,
                        style_assignment_a=None, style_assignment_b=None, digests=None,
-                       batch_id=None, recent_self=""):
+                       batch_id=None, recent_self="", lane=""):
     """DRAFT phase: assemble reddit ingredients, render via draft_prompt_core.
 
     The prompt TEXT lives in scripts/draft_prompt_core.py, the single source
@@ -1275,6 +1288,11 @@ def build_draft_prompt(project, config, candidates, top_report, recent_comments,
         "style_a_name": (style_assignment_a or {}).get("style") or "style_name",
         "style_b_name": (style_assignment_b or {}).get("style") or "style_name",
         "recent_self_block": recent_self or "",
+        # personal_brand renders the persona directive + persona-whitelisted
+        # project JSON in draft_prompt_core; comes from the plan stamp (see
+        # _discover_iteration), NEVER from env; salvage plans predate this
+        # cycle's lane and must keep the promotion directive.
+        "lane": lane or "",
     }
     return _dpc.render_reddit_prompt(ing)
 
@@ -1992,15 +2010,24 @@ def _discover_iteration(args, config, reddit_username, already_picked):
     builds the query bank and runs reddit_tools.cmd_search directly; no Claude
     session). Uses `decisions` key for downstream-phase compatibility.
     """
-    if args.project:
+    # Force-inject (2026-07-17): --project (manual MCP runs) or S4L_FORCE_PROJECT
+    # (the personal_brand lane's env, set by s4l_mode.py via the wrapper's eval)
+    # bypasses pick_project entirely. This is the ONLY way a persona:true entry
+    # can be selected; pick_project structurally excludes personas. Mirrors
+    # run-twitter-cycle.sh's S4L_FORCE_PROJECT handling.
+    force_project = args.project or (os.environ.get("S4L_FORCE_PROJECT") or "").strip()
+    if force_project:
         project = None
         for p in config.get("projects", []):
-            if p["name"].lower() == args.project.lower():
+            if p["name"].lower() == force_project.lower():
                 project = p
                 break
         if not project:
-            print(f"[post_reddit] ERROR: project '{args.project}' not found")
+            print(f"[post_reddit] ERROR: project '{force_project}' not found")
             return None
+        if not args.project:
+            print(f"[post_reddit] force-inject: S4L_FORCE_PROJECT={force_project} "
+                  f"(lane={os.environ.get('S4L_ACTIVE_LANE') or 'promotion'})")
     else:
         project = pick_project("reddit", exclude=already_picked)
         if not project:
@@ -2259,8 +2286,20 @@ def _discover_iteration(args, config, reddit_username, already_picked):
     # batch_id rides the plan file into --phase draft (snapshot filename,
     # top_performers --invoked-by, proposed_excludes activation gate). Same
     # reason as the salvage plan's batch_id above.
+    #
+    # lane (2026-07-17): stamped ONCE here at plan-write time and read from
+    # the PLAN by every downstream phase (_draft_iteration →
+    # render_reddit_prompt), never from env; salvage plans from earlier
+    # cycles and the env-less MCP approval poster must not inherit this
+    # cycle's lane. Same stamp-at-source convention as active_experiments
+    # (see CLAUDE.md). Derived from CONFIG (persona:true), not from
+    # S4L_ACTIVE_LANE, so an env-less manual `--project <persona>` run also
+    # drafts under the persona directive; the personal_brand lane can only
+    # ever force-inject the persona project, so the two always agree.
     return {"project_name": project_name, "decisions": selected,
             "batch_id": queue_batch,
+            "lane": ("personal_brand" if project.get("persona") is True
+                     else "promotion"),
             "cost": 0.0, "session_id": None,
             "phase": "discover"}
 
@@ -2434,7 +2473,8 @@ def _draft_iteration(plan, config, reddit_username):
                                 style_assignment_a=style_assignment_a,
                                 style_assignment_b=style_assignment_b, digests=digests,
                                 batch_id=plan.get("batch_id"),
-                                recent_self=_recent_self)
+                                recent_self=_recent_self,
+                                lane=plan.get("lane") or "")
     # Durable full-prompt record per batch (same prep-prompts dir the X
     # cycle writes; reddit-draft-prompt-<batch>.md, newest 50 kept), so
     # prompt-block presence is verifiable after any release.
@@ -2580,7 +2620,7 @@ def _draft_iteration(plan, config, reddit_username):
             merged_d["thread_title"] = drafted_d.get("thread_title") or c.get("thread_title")
             merged_d["engagement_style"] = _style_a or c.get("engagement_style")
             # Top-level assigned_style/assigned_mode default to Draft A (the
-            # legacy-mirror draft); the MCP post_drafts edit path overwrites
+            # legacy-mirror draft); the MCP approve_drafts edit path overwrites
             # these on the card when a human switches to Draft B (mirrors
             # twitter_post_plan.py's per-candidate override, see
             # _post_iteration below and index.ts's reddit approval branch).
@@ -2654,6 +2694,23 @@ def _post_iteration(plan, reddit_username):
         os.environ["CLAUDE_SESSION_ID"] = plan_session_id
 
     active_campaigns = load_active_reddit_campaigns()
+    # Persona posts never carry a product campaign suffix (2026-07-17): the
+    # reddit analog of the X lane's TWITTER_TAIL_LINK_RATE=0. Resolved from
+    # config (persona:true on the plan's project), NOT from lane env; the
+    # MCP approval poster runs env-less, and config is the durable truth.
+    if active_campaigns:
+        _plan_proj = (plan.get("project_name") or "").lower()
+        try:
+            _is_persona = any(
+                p.get("persona") is True and (p.get("name") or "").lower() == _plan_proj
+                for p in (load_config().get("projects") or [])
+            )
+        except Exception:
+            _is_persona = False
+        if _is_persona:
+            print(f"[post_reddit] persona project {_plan_proj!r}: campaign "
+                  f"suffixes disabled for this plan")
+            active_campaigns = []
     if active_campaigns:
         for c in active_campaigns:
             print(f"[post_reddit] active campaign id={c['id']} "
@@ -2681,7 +2738,7 @@ def _post_iteration(plan, reddit_username):
         # Two-draft cards (2026-07-15): each decision may carry its OWN
         # (assigned_style, assigned_mode) reflecting whichever draft is
         # actually posting — either Draft A (stamped at plan-write time) or
-        # Draft B (stamped by the MCP post_drafts edit path when a human
+        # Draft B (stamped by the MCP approve_drafts edit path when a human
         # switched drafts on the review card). Without this, every card
         # would coerce back to the single batch-wide Draft A assignment even
         # when it posted under Draft B's style, silently corrupting the

package/scripts/twitter_gen_links.py

@@ -148,7 +148,7 @@ def _tail_link_rate() -> float:
     # that calls this script). Force rate=1.0 there so a hand-approved draft
     # never drops the link the user already saw baked into the card text —
     # ported from the old post-time override in mcp/src/index.ts
-    # (`TWITTER_TAIL_LINK_RATE: "1.0"` on the post_drafts path), which no
+    # (`TWITTER_TAIL_LINK_RATE: "1.0"` on the approve_drafts path), which no
     # longer fires now that the decision happens here, before the card is
     # ever shown. The autonomous DRAFT_ONLY=0 lane keeps running the real A/B
     # experiment at the configured rate.
@@ -169,7 +169,7 @@ def apply_tail_link(candidate: dict, link_url: str) -> None:
     gate (both the review-card path and the autonomous-post path pass through
     twitter_gen_links.py first) and already budgets long waits (SEO page-gen
     alone can take 10-40 min). twitter_post_plan.py's post-time call was a bad
-    fit for the queue-backed pipeline: post_drafts is a synchronous MCP call
+    fit for the queue-backed pipeline: approve_drafts is a synchronous MCP call
     the user is actively waiting on after clicking Approve, while the
     s4l-worker scheduled task claims one job per minute and doesn't overlap a
     multi-minute drafting turn — so a queue-routed call there could stall an
@@ -313,7 +313,7 @@ def resolve_link(candidate: dict, projects: dict, page_gen_rate: float) -> tuple
     # `website`/`url` a persona project happens to carry (some installs got the
     # user's own X profile written there) must NEVER become a tail link. Enforce
     # it here at the single source so no downstream surface (review card, manual
-    # post_drafts) has to strip a link that should never have been generated.
+    # approve_drafts) has to strip a link that should never have been generated.
     if proj.get("persona"):
         return ("", "persona_no_link")
     plain_url = proj.get("website") or proj.get("url") or ""

package/scripts/twitter_post_plan.py

@@ -80,7 +80,7 @@ def print(*args, **kwargs):  # noqa: A001 -- deliberate builtins.print shadow
         _neuter_stream(sys.stderr)
 
 
-# Graceful SIGTERM: the MCP post_drafts timeout (and anything else that asks
+# Graceful SIGTERM: the MCP approve_drafts timeout (and anything else that asks
 # nicely before SIGKILL) sends SIGTERM. Dying instantly reopens the same
 # posted-but-unrecorded window as the broken pipe, so instead flag the loop to
 # stop at the next candidate boundary: current candidate finishes its
@@ -617,7 +617,7 @@ def post_one(c: dict, picker_assignment: dict | None = None) -> tuple[str, str]:
     # BEFORE the DRAFT_ONLY gate for every candidate, so both the review-card
     # path and the autonomous-post path already carry a stamped
     # tail_link_variant + finalized reply_text by the time they reach here.
-    # Post time is a bad fit for the queue-backed Claude call: post_drafts is
+    # Post time is a bad fit for the queue-backed Claude call: approve_drafts is
     # a synchronous MCP call the user is actively waiting on after clicking
     # Approve, while the s4l-worker scheduled task claims one job per minute
     # and doesn't overlap a multi-minute drafting turn, so a queue-routed call
@@ -1127,7 +1127,7 @@ def main() -> int:
 
     posted = skipped = failed = 0
     # Per-candidate outcome rows for the summary JSON. The MCP layer
-    # (mcp/src/index.ts post_drafts) prefers summary.candidate_results over
+    # (mcp/src/index.ts approve_drafts) prefers summary.candidate_results over
     # regex-parsing our stdout; this is the authoritative channel that marks
     # review-queue cards posted/terminal so a card is never re-fed to a poster
     # after its reply already landed (2026-07-06 duplicate-retry incident).
@@ -1262,7 +1262,7 @@ def main() -> int:
                 # (assigned_style, assigned_mode) reflecting whichever draft is
                 # actually posting, either the recommended one (stamped at
                 # plan-write time) or the other one (stamped by the MCP
-                # post_drafts edits path when a human switched drafts on the
+                # approve_drafts edits path when a human switched drafts on the
                 # review card). Without this, every card would coerce back to
                 # ONE cycle-wide style even when it posted under the other
                 # style, silently corrupting the engagement_style label the

package/setup/SKILL.md

@@ -71,7 +71,7 @@ Do not report setup complete until all of these are true:
 ## Choose the path
 
 - If the social-autoposter MCP tools are connected (`project_config`, `runtime`,
-  `queue_setup`, `post_drafts`, `get_stats`, `dashboard`), use the MCP path.
+  `queue_setup`, `approve_drafts`, `get_stats`, `dashboard`), use the MCP path.
   Do not hand-edit `config.json`.
 - If only the CLI/skill is installed, use the CLI fallback.
 

package/skill/run-reddit-search.sh

@@ -210,12 +210,21 @@ trap '_sa_emit_run_summary_oneshot; _sa_release_lease_oneshot; _sa_release_locks
 BATCH_ID="rdcycle-$(date +%Y%m%d-%H%M%S)"
 log "Cycle batch_id=$BATCH_ID"
 
+# --- Lane env (2026-07-17, mirrors run-draft-and-publish.sh) ----------------
+# personal_brand lane force-injects the persona project into the discover
+# phase (S4L_FORCE_PROJECT + S4L_ACTIVE_LANE=personal_brand); promotion-only
+# exports S4L_CYCLE_LANE=promotion and nothing else (normal weighted pick).
+# Both lanes on = weighted per-cycle coin flip inside s4l_mode.py, same as X.
+# Discover STAMPS the lane onto the plan JSON (post_reddit.py), so the draft
+# phase and the env-less MCP approval poster read the plan, never this env.
+eval "$(python3 "$REPO_DIR/scripts/s4l_mode.py" env 2>/dev/null || true)"
+
 # --- Draft-only mode (2026-07-14, mirrors run-draft-and-publish.sh) ---------
 # The ONE mode flag (scripts/s4l_mode.py draft-only, stdout 1/0) now governs
 # Reddit exactly like the X cycle: when ON, both lanes stop after the draft
 # phase and merge their drafted decisions into the menu-bar review cards
 # (merge_review_queue.py --reddit-plan); posting then happens one approved
-# card at a time via the plugin's post_drafts tool, which reconstructs a
+# card at a time via the plugin's approve_drafts tool, which reconstructs a
 # one-decision plan and reuses `post_reddit.py --phase post` unchanged. When
 # OFF (operator opt-out), the lanes post autonomously below as always.
 # Fail-safe default is 1 (draft-only) — same posture as the X wrapper.

package/skill/run-twitter-cycle.sh

@@ -1741,7 +1741,7 @@ fi
 # scan_candidates tool reads this so a Claude Desktop scheduled-task session can do
 # the drafting ITSELF (on the user's plan, no `claude -p`) and hand the drafts back
 # via submit_drafts. Candidates stay 'pending' (drafted+posted via submit_drafts ->
-# post_drafts, or salvaged by a later cycle). The browser lock was already released
+# approve_drafts, or salvaged by a later cycle). The browser lock was already released
 # at the Phase 1 handoff, so this exits clean via the EXIT trap. NO current caller
 # sets SCAN_ONLY, so the autopilot/draft_cycle paths are byte-for-byte unchanged.
 if [ "${SCAN_ONLY:-0}" = "1" ]; then