Update topic.md

.pi/agents/topic.md

@@ -49,9 +49,12 @@ Verify:
 ## 2. Load Context
 - Read {PROJECT_PATH}/icp.md
 - Read /workspace/content/images/images.json
-- List posts:
+- List candidate topic files (FILENAMES ONLY, do NOT read their contents):
   bash: ls /workspace/content/posts/*.md
 
+HARD RULE: At this stage you must not `read` any file inside
+`/workspace/content/posts/`. You only have the list of filenames.
+
 ---
 
 ## 3. Filter Recent Topics
@@ -65,24 +68,35 @@ Verify:
 
 ---
 
-## 4. Score Candidates
+## 4. Score Candidates (FILENAME ONLY)
 
-Do not open all content markdown files. Decide on the filename.
+HARD RULE: Do NOT open or `read` any candidate markdown file in this step.
+The filename IS the headline — decide purely from the filename tokens.
 
-Score each filename:
-- +1 keyword match (from icp.md)
-- +1 pain point match
+For each candidate filename:
+1. Strip the `.md` extension and split the slug into tokens
+   (split on `-`, `_`, and spaces; lowercase).
+2. Score:
+   - +1 for each token that matches a keyword from `icp.md`
+   - +1 for each token that matches a pain point from `icp.md`
+3. Pick the highest score. Tie-break alphabetically by filename.
 
-Pick highest score (tie → alphabetical) and select this markdown file
+Output of this step: exactly ONE selected filename.
 
 ---
 
 ## 5. Generate Output
-- Read only the selected markdown file for content. Do not read all content files
+
+HARD RULE: In this entire step you may `read` exactly ONE markdown file —
+the single filename selected in step 4. Reading any other post file is a
+protocol violation. If you catch yourself about to read a second post
+file, stop and use only what you already have.
+
+- Read the selected markdown file (and only that one).
 - Extract:
   - Story bullets
   - Pain points (max 3)
-- Pick matching image from images.json
+- Pick matching image from images.json (already loaded in step 2).
 
 Create:
 - Headline (Title Case filename)