@m13v/s4l 1.7.6-rc.8 → 1.7.6

package/scripts/draft_prompt_core.py

@@ -76,6 +76,12 @@ def _config_path():
 ARM_TREATMENT = "treatment_v4"
 ARM_CONTROL = "control_v4"
 
+# The "treatment_v4 assigns no real style" decision (and its sentinel value)
+# lives ONLY in engagement_styles.py now (STYLE_SENTINEL_TREATMENT,
+# pick_style_for_post, pick_exploration_style, 2026-07-17 consolidation):
+# every platform's picker routes through those two functions, so this
+# module never needs its own copy of the sentinel or the arm check for it.
+
 
 def pick_draft_prompt_arm(env=None, export=True):
     """Return (variant, source). source in {env, pin, coin, fallback}.
@@ -118,17 +124,84 @@ def pick_draft_prompt_arm(env=None, export=True):
 # Text transcribed verbatim from run-twitter-cycle.sh (2026-07-16).
 # --------------------------------------------------------------------------
 
-_DIRECTIVE_TREATMENT = "Otherwise: draft a direct, natural reply that stands on its own as a useful contribution to the thread. Mention the matched project only when it is genuinely the most relevant thing to say, and state it plainly in one clause; most replies will not need it. PRIORITY ORDER for how you write this, highest first: (1) learned_preferences.draft_style_notes and edit_examples under PROJECT ROUTING are the strongest signal available, real human corrections to this account's own past drafts, and are MANDATORY, not advisory. (2) The ACCOUNT VOICE CORPUS block and the matched project's voice.examples are VERBATIM GROUND TRUTH for how this account actually writes: capitalization, punctuation, contractions, sentence length, terseness. Match those mechanics exactly. (3) voice.tone is a supporting description only, never ground truth: if it ever conflicts with what the corpus/examples actually show, follow the examples, not the tone description. (4) The assigned engagement style is an optional idea for angle and length, not a required structure; draw on it only when it fits naturally, never at the cost of (1)-(3). Do NOT use the concede-then-reverse skeleton in ANY form. Banned openings include: 'X is the easy part/half/win, the hard part is Y'; 'X was never the [thing], it's Y'; 'X isn't the [problem], it's Y'; 'the real/actual/harder part is Y'; 'what actually breaks/ships/matters is Y'; 'the part nobody says/shows is Y'; 'X is solved, Y is what breaks'. Lead with substance from ONE entry point and vary the entry point across replies: a concrete first-hand specific or number; a direct answer to the exact question asked; one sharp opinion with no hedge; a genuine question that moves the thread forward; or a relevant pointer. No warm-up framing sentence before the substance. Length is governed by the per-style LENGTH LIMIT in the style block above. NEVER em dashes. Never violate voice.never. Treat learned_preferences.audience_avoid / thread_avoid matches as strong reasons to skip the candidate. Never violate content_guardrails.do_not."
+# --------------------------------------------------------------------------
+# Shared treatment_v4 core (2026-07-17). Written ONCE, used by BOTH lanes.
+#
+# Why this exists: promotion and persona used to carry independent copies of
+# nearly-identical priority-order/skeleton-ban/vary-entry-point text. They
+# drifted -- persona's copy was missing the "vary the entry point across
+# replies" sentence promotion had, which a real batch comparison traced to
+# the "nobody ___s it" tell appearing 2x as often in treatment as control:
+# a batch shares ONE assigned style, and without an explicit
+# vary-across-the-batch instruction the model reused that style's defining
+# vocabulary verbatim across every unrelated candidate. Do NOT fork this
+# text again for a lane-specific tweak; parameterize _treatment_core()
+# instead, or add a lane-specific sentence OUTSIDE it.
+# --------------------------------------------------------------------------
 
-_DIRECTIVE_CONTROL = "Otherwise: draft a reply using the best engagement style. Length is governed ENTIRELY by the per-style LENGTH LIMIT in the style block above; obey that target and ceiling, do not apply any other length rule here. NEVER em dashes. Apply the matched project's `voice` block from ALL_PROJECTS_JSON: follow voice.tone, never violate voice.never, mirror voice.examples / voice.examples_good when present. The global learned_preferences block under PROJECT ROUTING is distilled human review feedback and is MANDATORY, not advisory: follow every learned_preferences.draft_style_notes entry when writing (it overrides the engagement style's structural template on conflict), and treat learned_preferences.audience_avoid / thread_avoid matches as strong reasons to skip the candidate. Never violate content_guardrails.do_not."
+def _treatment_core(voice_examples_ref):
+    """Shared PRIORITY ORDER + skeleton ban + entry-point-variation text.
+
+    voice_examples_ref: the lane-specific noun phrase for priority item (2)
+    (promotion scopes to "the matched project's voice.examples"; persona
+    scopes to "voice.examples above" since it's already narrowed to the one
+    persona project). No mention of engagement style anywhere in here on
+    purpose: treatment_v4 assigns none (see engagement_styles.py's
+    STYLE_SENTINEL_TREATMENT / pick_style_for_post).
+    """
+    return (
+        "PRIORITY ORDER for how you write this, highest first: (1) "
+        "learned_preferences.draft_style_notes are the "
+        "strongest signal available, real human corrections to this "
+        "account's own past drafts (some carry a before/after example), and "
+        "are MANDATORY, not advisory. (2) "
+        f"The ACCOUNT VOICE CORPUS block and {voice_examples_ref} are "
+        "VERBATIM GROUND TRUTH for how this account actually writes: "
+        "capitalization, punctuation, contractions, sentence length, "
+        "terseness. Match those mechanics exactly. (3) voice.tone is a "
+        "supporting description only, never ground truth: if it ever "
+        "conflicts with what the corpus/examples actually show, follow the "
+        "examples, not the tone description. Do NOT use the "
+        "concede-then-reverse skeleton in ANY form. Banned openings "
+        "include: 'X is the easy part/half/win, the hard part is Y'; 'X "
+        "was never the [thing], it's Y'; 'X isn't the [problem], it's Y'; "
+        "'the real/actual/harder part is Y'; 'what actually "
+        "breaks/ships/matters is Y'; 'the part nobody says/shows is Y'; 'X "
+        "is solved, Y is what breaks'. Lead with substance from ONE entry "
+        "point and vary the entry point AND the rhetorical move across "
+        "every reply you draft this batch, even when the same topic or "
+        "style would naturally suggest reusing the same construction: a "
+        "concrete first-hand specific or number; a direct answer to the "
+        "exact question asked; one sharp opinion with no hedge; a genuine "
+        "question that moves the thread forward; or a relevant pointer. No "
+        "warm-up framing sentence before the substance."
+    )
 
-_DIVERGENCE_NOTE_TREATMENT = " Beyond following different style templates: draft_a_text and draft_b_text must be genuinely distinct takes on this thread -- think of them as variant 1a and variant 1b: different entry point, different specific detail seized on, different rhetorical move -- not two phrasings of the same underlying observation dressed in different style language. If 1a and 1b would say essentially the same thing, pick a different angle entirely for 1b rather than just rewording 1a."
 
-_PERSONA_SKELETON_BAN = " Also do NOT use the concede-then-reverse skeleton in ANY form: banned openings include 'X is the easy part/half/win, the hard part is Y', 'X was never the [thing], it's Y', 'X isn't the [problem], it's Y', 'the real/actual/harder part is Y', 'what actually breaks/ships/matters is Y', 'the part nobody says/shows is Y', and 'X is solved, Y is what breaks'; if a draft has that pivot, rewrite it from one of the entry points above."
+_TREATMENT_LENGTH_NOTE = (
+    "Length: match how long the account's own real examples above actually "
+    "run; when the corpus doesn't make it obvious, default to one or two "
+    "sentences, well under Twitter's 250-character practical limit. NEVER "
+    "em dashes."
+)
+
+_DIRECTIVE_TREATMENT = (
+    "Otherwise: draft a direct, natural reply that stands on its own as a "
+    "useful contribution to the thread. Mention the matched project only "
+    "when it is genuinely the most relevant thing to say, and state it "
+    "plainly in one clause; most replies will not need it. "
+    + _treatment_core("the matched project's voice.examples") +
+    " " + _TREATMENT_LENGTH_NOTE +
+    " Never violate voice.never. Treat learned_preferences.audience_avoid "
+    "/ thread_avoid matches as strong reasons to skip the candidate. Never "
+    "violate content_guardrails.do_not."
+)
 
-_PERSONA_VOICE_PRIORITY_CLAUSE = " PRIORITY ORDER for how you write this, highest first: (1) learned_preferences.draft_style_notes and edit_examples are the strongest signal available, real human corrections to this account's own past drafts, and are MANDATORY. (2) The ACCOUNT VOICE CORPUS block and voice.examples above are VERBATIM GROUND TRUTH for how they actually write: capitalization, punctuation, contractions, sentence length, terseness. Match those mechanics exactly. (3) voice.tone is a supporting description only, never ground truth: if it ever conflicts with what the corpus/examples actually show (for example tone says lowercase is fine but every real example is properly capitalized), follow the examples. (4) The assigned engagement style is an optional idea for angle and length, not a required structure."
+_DIRECTIVE_CONTROL = "Otherwise: draft a reply using the best engagement style. Length is governed ENTIRELY by the per-style LENGTH LIMIT in the style block above; obey that target and ceiling, do not apply any other length rule here. NEVER em dashes. Apply the matched project's `voice` block from ALL_PROJECTS_JSON: follow voice.tone, never violate voice.never, mirror voice.examples / voice.examples_good when present. The global learned_preferences block under PROJECT ROUTING is distilled human review feedback and is MANDATORY, not advisory: follow every learned_preferences.draft_style_notes entry when writing (it overrides the engagement style's structural template on conflict), and treat learned_preferences.audience_avoid / thread_avoid matches as strong reasons to skip the candidate. Never violate content_guardrails.do_not."
 
-_PERSONA_DIRECTIVE_TEMPLATE = "Otherwise: draft a reply that stands on its own as a genuinely useful contribution to THIS thread. Ground it in the persona's real, first-hand experience from the ACCOUNT VOICE CORPUS block below (specific projects, real numbers, sharp opinions, actual failures) and in the persona's `voice` block from ALL_PROJECTS_JSON. Add exactly ONE of: a concrete specific from that lived experience, a sharp non-obvious opinion, a useful pointer, or a question that genuinely moves the thread forward. NEVER generic agreement ('makes sense', 'this is spot on', 'great point', 'the nuance here is').@SKELETON_BAN@ This is a personal account, not a brand: sound like a real person in the thread. If web search is available and the thread hinges on a current fact, verify it before drafting rather than guessing. Length is governed by the per-style LENGTH LIMIT in the style block above. NEVER em dashes. Follow voice.tone, never violate voice.never, mirror voice.examples / voice.examples_good when present.@VOICE_PRIORITY@ The global learned_preferences block under PROJECT ROUTING is distilled human review feedback and is MANDATORY, not advisory: follow every learned_preferences.draft_style_notes entry when writing @PREFS_RELATION@, and treat learned_preferences.audience_avoid / thread_avoid matches as strong reasons to skip the candidate. Never violate content_guardrails.do_not."
+_DIVERGENCE_NOTE_TREATMENT = " draft_a_text and draft_b_text must be genuinely distinct takes on this thread regardless of style -- think of them as variant 1a and variant 1b: different entry point, different specific detail seized on, different rhetorical move -- not two phrasings of the same underlying observation. If 1a and 1b would say essentially the same thing, pick a different angle entirely for 1b rather than just rewording 1a."
+
+_PERSONA_DIRECTIVE_TEMPLATE = "Otherwise: draft a reply that stands on its own as a genuinely useful contribution to THIS thread. Ground it in the persona's real, first-hand experience from the ACCOUNT VOICE CORPUS block below (specific projects, real numbers, sharp opinions, actual failures) and in the persona's `voice` block from ALL_PROJECTS_JSON. Add exactly ONE of: a concrete specific from that lived experience, a sharp non-obvious opinion, a useful pointer, or a question that genuinely moves the thread forward. NEVER generic agreement ('makes sense', 'this is spot on', 'great point', 'the nuance here is').@TREATMENT_CORE@ This is a personal account, not a brand: sound like a real person in the thread. If web search is available and the thread hinges on a current fact, verify it before drafting rather than guessing.@LENGTH_NOTE@ Follow voice.tone, never violate voice.never, mirror voice.examples / voice.examples_good when present. The global learned_preferences block under PROJECT ROUTING is distilled human review feedback and is MANDATORY, not advisory: follow every learned_preferences.draft_style_notes entry when writing @PREFS_RELATION@, and treat learned_preferences.audience_avoid / thread_avoid matches as strong reasons to skip the candidate. Never violate content_guardrails.do_not."
 
 
 def draft_directive(arm=None, lane=None):
@@ -137,17 +210,17 @@ def draft_directive(arm=None, lane=None):
     lane = lane if lane is not None else os.environ.get("S4L_ACTIVE_LANE", "")
     if lane == "personal_brand":
         if arm == ARM_TREATMENT:
-            ban = _PERSONA_SKELETON_BAN
-            priority = _PERSONA_VOICE_PRIORITY_CLAUSE
-            rel = "(per the PRIORITY ORDER above: learned_preferences and the account's real voice outrank the engagement style whenever they disagree)"
+            core = " " + _treatment_core("voice.examples above")
+            length_note = " " + _TREATMENT_LENGTH_NOTE
+            rel = "(per the PRIORITY ORDER above: learned_preferences and the account's real voice are what drive this reply)"
         else:
-            ban = ""
-            priority = ""
+            core = ""
+            length_note = " Length is governed by the per-style LENGTH LIMIT in the style block above. NEVER em dashes."
             rel = "(it overrides the engagement style's structural template on conflict)"
         return (
             _PERSONA_DIRECTIVE_TEMPLATE
-            .replace("@SKELETON_BAN@", ban)
-            .replace("@VOICE_PRIORITY@", priority)
+            .replace("@TREATMENT_CORE@", core)
+            .replace("@LENGTH_NOTE@", length_note)
             .replace("@PREFS_RELATION@", rel)
         )
     if arm == ARM_TREATMENT:
@@ -260,6 +333,9 @@ def global_learned_prefs_json():
         config = _load_config()
         block = lp.get_global_block(config)
         block.pop("history", None)
+        # edit_examples feed only the digest now (2026-07-17); the drafting and
+        # judging prompts read their distilled form via draft_style_notes.
+        block.pop("edit_examples", None)
         return json.dumps(block, indent=2)
     except Exception:
         return "{}"
@@ -279,24 +355,6 @@ def recent_self_block(platform, limit=20):
         return ""
 
 
-def product_names():
-    """Every configured project name (lowercased) plus the agency tag, for
-    the reddit never-name-the-product rule. Falls back to a generic list on
-    config failure so the rule never silently vanishes."""
-    names = []
-    try:
-        config = _load_config()
-        for p in config.get("projects", []):
-            n = (p.get("name") or "").strip().lower()
-            if n and n not in names:
-                names.append(n)
-    except Exception:
-        pass
-    if "s4l" not in names:
-        names.append("s4l")
-    return names
-
-
 # --------------------------------------------------------------------------
 # TWITTER renderer. Byte-exact transcription of the run-twitter-cycle.sh
 # PREP_PROMPT heredoc (2026-07-16). @TOKENS@ are unique and replaced below;
@@ -354,7 +412,7 @@ For each chosen candidate:
    - If the candidate block shows an EXISTING DRAFT line AND draft age < 30 minutes, REUSE the draft text verbatim as draft_a_text/draft_a_style (set is_reused_draft=true, draft_b_text=null, draft_b_style=null). Do NOT call log_draft.py; do NOT redraft; do NOT write a second variant, prior cycle already paid the LLM cost for the one draft you have.
    - Otherwise (fresh candidate, is_reused_draft=false): write TWO independent drafts. Do NOT judge, rank, or pick a favorite between them, both are shown to the reviewer, who decides.
      - draft_a_text: follow the DRAFT A style block above (its own description/example/note/length limit).
-     - draft_b_text: follow the DRAFT B style block above, written INDEPENDENTLY from scratch as if draft_a_text did not exist. Do NOT lightly reword draft_a_text into draft_b_text, they must diverge in length and rhetorical move because they follow different style templates, not just differ in phrasing. If you notice draft_b_text ending up as a paraphrase of draft_a_text, stop and rewrite it from Style B's own example instead.@DRAFT_B_DIVERGENCE_NOTE@
+     - draft_b_text: follow the DRAFT B style block above, written INDEPENDENTLY from scratch as if draft_a_text did not exist. Do NOT lightly reword draft_a_text into draft_b_text, they must diverge in length and rhetorical move, not just differ in phrasing. If you notice draft_b_text ending up as a paraphrase of draft_a_text, stop and rewrite it from Style B's own example instead.@DRAFT_B_DIVERGENCE_NOTE@
    - @DRAFT_DIRECTIVE@ (applies to both drafts on fresh candidates; each still obeys its OWN style's length limit, not a shared one).
 3a. PERSIST DRAFT A (skip entirely for reused drafts):
      python3 @REPO_DIR@/scripts/log_draft.py --candidate-id CANDIDATE_ID --text 'DRAFT_A_TEXT' --style DRAFT_A_STYLE --assigned-style '@PICKED_STYLE@' --assigned-mode '@PICKED_MODE@'
@@ -439,7 +497,18 @@ def render_twitter_prompt(ing):
     """
     arm = ing.get("arm") or os.environ.get("S4L_DRAFT_PROMPT_VARIANT") or ""
     lane = ing.get("lane") if ing.get("lane") is not None else os.environ.get("S4L_ACTIVE_LANE", "")
+    # picked_style/picked_style_b already carry the sentinel for treatment_v4
+    # (engagement_styles.pick_style_for_post/pick_exploration_style refuse to
+    # assign a real style under that arm, 2026-07-17 consolidation), no
+    # override needed here -- this used to re-derive the sentinel from `arm`
+    # independently of the picker, which is exactly the kind of duplicated
+    # decision that let a second, unfixed copy (Reddit's post_reddit.py,
+    # which calls the picker directly and never went through this function)
+    # silently keep assigning real styles under treatment_v4.
     picked_style = ing.get("picked_style") or ""
+    picked_style_b = ing.get("picked_style_b") or ""
+    picked_mode = ing.get("picked_mode") or "use"
+    picked_mode_b = ing.get("picked_mode_b") or "use"
     return (
         _TW_TEMPLATE
         .replace("@PREFIX@", ing.get("prefix") or "")
@@ -459,10 +528,10 @@ def render_twitter_prompt(ing):
         .replace("@DRAFT_B_DIVERGENCE_NOTE@", draft_b_divergence_note(arm))
         .replace("@DRAFT_DIRECTIVE@", draft_directive(arm, lane))
         .replace("@PICKED_STYLE_OR_INVENT@", picked_style or "(invent)")
-        .replace("@PICKED_STYLE_B@", ing.get("picked_style_b") or "")
-        .replace("@PICKED_MODE_B@", ing.get("picked_mode_b") or "use")
+        .replace("@PICKED_STYLE_B@", picked_style_b)
+        .replace("@PICKED_MODE_B@", picked_mode_b)
         .replace("@PICKED_STYLE@", picked_style)
-        .replace("@PICKED_MODE@", ing.get("picked_mode") or "use")
+        .replace("@PICKED_MODE@", picked_mode)
     )
 
 
@@ -526,21 +595,6 @@ OMIT THESE (clear no-bridge cases only):
 ## THREAD CONTENT (pre-fetched)
 Each candidate above carries a THREAD CONTENT block (OP + top comments), already fetched for you. Never fetch, navigate, or open any reddit.com URL and never try to reload a thread; apply the SELECTION GATE using the inlined content only. WebSearch and WebFetch ARE available for EXTERNAL fact-checking: use them ONLY when a thread hinges on a current fact, a name, a release, or a claim you are not sure about, so your comment is specific and correct instead of vague. Most comments need no search at all. If a candidate is somehow missing its THREAD CONTENT block, OMIT it.
 
-## CRITICAL CONTENT RULES (apply only to threads that pass the gate)
-These platform rules are ABSOLUTE and override anything more permissive in the DRAFT DIRECTIVE below (including any allowance to mention the project).
-- Go BIMODAL on length: 1 punchy sentence (<100 chars) OR 4-5 sentences of real substance. Avoid 2-3 sentence middle-ground.
-- GROUNDING RULE — pick ONE lane per comment:
-  LANE 1 - DISCLOSED STORY: open with a hedge ("hypothetically", "imagine someone running this", "scenario:") then you may invent specifics freely.
-  LANE 2 - NO FABRICATION: every specific (numbers, durations, places, tools) must appear verbatim in the content_angle above. Otherwise drop the specific and pattern-frame ("the part that breaks down is...", "the typical failure mode is...").
-- VOICE RELATIONSHIP: see the dedicated section below; it governs whether you speak AS the maker or as an outside observer.
-- NEVER mention product names (@PRODUCT_NAMES@).
-- NEVER include URLs or links in your comment text.
-- Prefer replying to OP (top-level reply). ONE comment per thread.
-- Statements beat questions. Be authoritative, not inquisitive.
-
-## Content rules
-@CONTENT_RULES@
-
 ## DRAFT A: assigned style
 @STYLES_BLOCK_A@
 
@@ -550,15 +604,13 @@ These platform rules are ABSOLUTE and override anything more permissive in the D
 ## DRAFT DIRECTIVE
 @DRAFT_DIRECTIVE@ (applies to both drafts; each still obeys its OWN style's length limit, not a shared one.)
 
-@VOICE_RELATIONSHIP@
-
 ## PERSIST EACH DRAFT (keeps this session alive; do it per thread, not at the end)
 Right after drafting BOTH texts for a thread, persist Draft A (never Draft B) with ONE quick Bash call, then move to the next thread:
      python3 @REPO_DIR@/scripts/log_draft.py --platform reddit --thread-url 'THREAD_URL' --text 'DRAFT_A_TEXT' --style DRAFT_A_STYLE
 Work ONE thread at a time: gate it, draft both texts, persist, next. Failure here is non-fatal, log a warning and continue. Only after EVERY thread is handled do you assemble and return the single result JSON.
 
 ## OUTPUT FORMAT
-Return ONE JSON object with two arrays (a JSON schema is enforced on this session). Both draft_a_text and draft_b_text are REQUIRED for every posts[] entry — write both, under their respective assigned styles above, applying the CRITICAL CONTENT RULES and Content rules to each independently:
+Return ONE JSON object with two arrays (a JSON schema is enforced on this session). Both draft_a_text and draft_b_text are REQUIRED for every posts[] entry — write both, under their respective assigned styles above, applying the DRAFT DIRECTIVE to each independently:
 
 {"posts": [...], "rejects": [...]}
 
@@ -593,6 +645,19 @@ Examples of WRONG proposals (do not emit):
 Do NOT narrate beyond the persist calls. Gate, draft-or-reject, persist, return the single JSON object."""
 
 
+def _strip_grounding_section(styles_block):
+    """Drop the '## GROUNDING RULE' tail that get_assigned_style_prompt
+    appends to every style block. Reddit-draft-only (2026-07-18 content-rule
+    deprecation): the reddit draft prompt no longer carries the GROUNDING
+    RULE or VOICE RELATIONSHIP sections, so the style-block copy would both
+    resurrect a deprecated rule and dangle a reference to a section that no
+    longer exists. X and the engage rails keep their copies (their prompts
+    still render the sections it points at). Safe as truncation because the
+    grounding rule is the LAST thing get_assigned_style_prompt appends."""
+    i = (styles_block or "").find("## GROUNDING RULE")
+    return styles_block[:i].rstrip() if i != -1 else (styles_block or "")
+
+
 def render_reddit_prompt(ing):
     """Render the full Reddit draft prompt. Keys in `ing`:
     project_name, content_angle, candidates_block, n_candidates, batch_id,
@@ -601,14 +666,16 @@ 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
-
     top_ctx = ""
     if not skip_top_report(arm):
         top_report = ing.get("top_report") or ""
@@ -630,15 +697,16 @@ def render_reddit_prompt(ing):
         .replace("@BATCH_ID@", ing.get("batch_id") or "")
         .replace("@RECENT_SELF_BLOCK@", ing.get("recent_self_block") or "")
         .replace("@TOP_CTX@", top_ctx)
-        .replace("@PRODUCT_NAMES@", ", ".join(product_names()))
-        .replace("@CONTENT_RULES@", get_content_rules("reddit"))
-        .replace("@STYLES_BLOCK_A@", ing.get("styles_block_a") or "")
-        .replace("@STYLES_BLOCK_B@", ing.get("styles_block_b") or "")
+        .replace("@STYLES_BLOCK_A@", _strip_grounding_section(ing.get("styles_block_a")))
+        .replace("@STYLES_BLOCK_B@", _strip_grounding_section(ing.get("styles_block_b")))
         .replace("@DRAFT_DIRECTIVE@", draft_directive(arm, lane))
-        .replace("@VOICE_RELATIONSHIP@", get_voice_relationship_rule())
         .replace("@STYLE_A_NAME@", ing.get("style_a_name") or "style_name")
         .replace("@STYLE_B_NAME@", ing.get("style_b_name") or "style_name")
         .replace("@PROJECT_NAME@", project_name)
+        # The treatment length note is shared text written for X; on Reddit
+        # keep the same short-length instruction without the platform label.
+        .replace("well under Twitter's 250-character practical limit",
+                 "well under 250 characters")
     )
 
 

package/scripts/engage_github.py

@@ -40,7 +40,13 @@ except Exception:  # never let a missing module break the engage lane
         return ""
 
 REPO_DIR = os.path.expanduser("~/social-autoposter")
-CONFIG_PATH = os.path.join(REPO_DIR, "config.json")
+# THE canonical config loader (scripts/config.py): S4L_CONFIG_PATH / state-dir /
+# S4L_REPO_DIR aware, mtime-cached. Replaces this file's hand-rolled loader and
+# its hardcoded config path (the S4L-4H dead-path class on customer boxes).
+import os as _cfg_os, sys as _cfg_sys
+_cfg_sys.path.insert(0, _cfg_os.path.dirname(_cfg_os.path.abspath(__file__)))
+from config import config_path as _canonical_config_path, load_config
+CONFIG_PATH = _canonical_config_path()
 REPLY_DB = os.path.join(REPO_DIR, "scripts", "reply_db.py")
 SKILL_FILE = os.path.join(REPO_DIR, "SKILL.md")
 
@@ -60,9 +66,6 @@ os.environ["S4L_PYTHON"] = PYTHON
 THREAD_CHAR_CAP = 12000
 
 
-def load_config():
-    with open(CONFIG_PATH) as f:
-        return json.load(f)
 
 
 def get_next_pending():
@@ -451,7 +454,10 @@ def main():
     # (HTTP-only now; per-account scoping is resolved inside the helper.)
     from github_tools import _dynamic_owner_blocklist
     excluded_repos = excluded_repos | _dynamic_owner_blocklist()
-    our_username = config.get("accounts", {}).get("github", {}).get("username", "m13v")
+    # Resolved identity, never a hardcoded default (the old "m13v" fallback
+    # silently self-filtered/attributed as the repo owner on other installs).
+    from account_resolver import resolve as _resolve_account
+    our_username = _resolve_account("github") or ""
 
     start_time = time.time()
     processed = 0

package/scripts/engage_reddit.py

@@ -27,7 +27,13 @@ sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
 from http_api import api_get, api_post
 
 REPO_DIR = os.path.expanduser("~/social-autoposter")
-CONFIG_PATH = os.path.join(REPO_DIR, "config.json")
+# THE canonical config loader (scripts/config.py): S4L_CONFIG_PATH / state-dir /
+# S4L_REPO_DIR aware, mtime-cached. Replaces this file's hand-rolled loader and
+# its hardcoded config path (the S4L-4H dead-path class on customer boxes).
+import os as _cfg_os, sys as _cfg_sys
+_cfg_sys.path.insert(0, _cfg_os.path.dirname(_cfg_os.path.abspath(__file__)))
+from config import config_path as _canonical_config_path, load_config
+CONFIG_PATH = _canonical_config_path()
 REPLY_DB = os.path.join(REPO_DIR, "scripts", "reply_db.py")
 CAMPAIGN_BUMP = os.path.join(REPO_DIR, "scripts", "campaign_bump.py")
 REDDIT_MCP_CONFIG = os.path.expanduser("~/.claude/browser-agent-configs/reddit-agent-mcp.json")
@@ -110,9 +116,6 @@ def _release_browser_lease() -> None:
         pass
 
 
-def load_config():
-    with open(CONFIG_PATH) as f:
-        return json.load(f)
 
 
 def load_active_reddit_campaigns():
@@ -386,7 +389,11 @@ def get_recent_archetypes(platform, limit=3):
 
 def build_prompt(reply, recent_replies, config, excluded_authors, top_report="", prior_history_block="", meta_callout=None):
     """Build a minimal prompt for one reply."""
-    reddit_username = config.get("accounts", {}).get("reddit", {}).get("username", "Deep_Ad1959")
+    # Resolve through the one account resolver (env -> config); never a hardcoded
+    # username. Empty means "unknown account" (the prompt just omits it) rather
+    # than silently impersonating the repo owner on a misconfigured install.
+    from account_resolver import resolve as _resolve_account
+    reddit_username = _resolve_account("reddit") or ""
     reply_json = json.dumps(reply, indent=2)
 
     # Moltbook: skip recent_replies + top_report context blocks. Both are

package/scripts/engagement_signal_feedback.py

@@ -55,7 +55,6 @@ MIN_STYLE_SAMPLE = 10
 WARM_MIN_LAST_30D = 4
 WARM_MIN_CHILD_REPLIES = 1
 MAX_WARM_PROMOTIONS_PER_RUN = 3
-ENTRY_CHAR_LIMIT = 200  # mirrors learned_preferences.MAX_ENTRY_CHARS
 
 
 def fetch_window_rows(platform: str, days: int, limit_pages: int = 20) -> list[dict]:
@@ -165,7 +164,7 @@ def build_warm_entry(w: dict) -> str:
         f"({w['replied']}/{w['last_30d']} replied in 30d, {w['child_replies']} child "
         f"replies, 0 risk skips); warm to more engaged/product-relevant replies."
     )
-    return entry[:ENTRY_CHAR_LIMIT]
+    return entry
 
 
 def build_style_note(top: dict) -> str:
@@ -174,8 +173,6 @@ def build_style_note(top: dict) -> str:
         f"({top['child_reply_rate']:.2f}/reply, n={top['n']}); prefer it for technical-"
         "correction replies (concrete metric, name the exact failure mode)."
     )
-    if len(note) > ENTRY_CHAR_LIMIT:
-        note = note[: ENTRY_CHAR_LIMIT - 1].rsplit(" ", 1)[0] + "…"
     return note