feat(ops-socials): auto-consume post-performance learnings before composing

[auroracapital]

What

/ops-socials now reads auto-generated post-performance learnings before composing any draft and biases tone/format toward what the data shows works.

• New section "Auto-consume performance learnings before composing" — reads $PREFS_PATH/social-metrics/learnings.md (ranked do-more / do-less features + top-performer templates), applies it, falls back to a house default when absent or COLLECTING.
• Pointer added to the "Draft a tweet" recipe so the step is unmissable.

Why

Closes the optimization loop. An always-on tracker (launchd, owner-local) pulls each Typefully social set's analytics on a schedule, writes a time series, and re-derives the learnings. This PR is the consume half: tracker measures live posts → updates learnings → skill biases the next draft → repeat. Tone/format self-improves from real engagement data instead of guesswork.

Safety

• No personal data. References the plugin data dir ($PREFS_PATH/social-metrics/) per Rule 0, never an owner-specific path/handle.
• tests/test-no-secrets.sh: 14 passed, 0 failed. Manual personal-data grep: clean.
• Docs-only change to one SKILL.md (+21/-1). No tool/permission changes.

🤖 Generated with Claude Code

---

Note

Low Risk
Documentation-only change to agent skill instructions; no runtime, permissions, or posting behavior changes in code.

Overview
/ops-socials now tells the agent to read and apply owner-local post-performance learnings before any personal/founder Typefully draft (single post, thread, or cross-platform)—not for upload-post project brands.

A new section defines a bash check on $PREFS_PATH/social-metrics/learnings.md: use ranked Do MORE / Do LESS and top-performer templates when the file exists and status is not COLLECTING; otherwise use a fixed house default voice. The agent must note in one line which learnings it applied. Learnings come from an external launchd tracker; this skill only consumes them.

The "Draft a tweet / Make this a thread" recipe now explicitly starts with that consume step before typefully_create_draft.

^{Reviewed by Cursor Bugbot for commit aab2f42. Bugbot is set up for automated code reviews on this repo. Configure here.}

Summary by CodeRabbit

• New Features

  • Personal social post drafts now automatically incorporate performance learnings, biasing tone and format toward successful content patterns when available.

• Documentation

  • Updated tweet and thread drafting recipes to reflect automatic performance learnings integration in the drafting workflow.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 4c89bd11-c835-42e2-9f7c-5aba37ef5465

📥 Commits

Reviewing files that changed from the base of the PR and between 00a0f47 and aab2f42.

📒 Files selected for processing (1)

• claude-ops/skills/ops-socials/SKILL.md

---

📝 Walkthrough

Walkthrough

Updated the ops-socials skill documentation to add conditional performance learnings consumption before drafting personal Typefully posts. A new guidance section describes reading learnings when available and applying performance biases, while the tweet/thread recipe is updated to reference this guidance.

Changes

Performance Learnings Integration

Layer / File(s)

Summary

Auto-consume performance learnings guidance claude-ops/skills/ops-socials/SKILL.md

Introduces a new subsection that conditionally reads $PREFS_PATH/social-metrics/learnings.md for personal/founder Typefully drafting when present and not in COLLECTING status, instructing bias toward "Do MORE of" features and top-performer templates, with fallback to default voice formatting. Updates the "Draft a tweet / Make this a thread" recipe to reference and apply the learnings before staging the draft.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~5 minutes

Possibly related PRs

• Lifecycle-Innovations-Limited/claude-ops#335: Extended the existing ops-socials router documentation by adding instructions to auto-consume social-metrics/learnings.md before Typefully tweet/thread drafting, building on the router's earlier draft/Typefully staging recipes.

Poem

🐰 With whiskers twitching, hop and bound,
Performance learnings now abound!
Tweets shall flow with better zest,
Learning what works best, then all the rest. 🚀

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately reflects the main change: auto-consuming post-performance learnings before composing social posts, which is the core objective documented in the PR.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch feat/ops-socials-learnings-autoconsume

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[cursor]

Cursor Bugbot has reviewed your changes and found 2 potential issues.

Bugbot Autofix prepared fixes for both issues found in the latest run.

• ✅ Fixed: COLLECTING status contradicts file-exists rule
  • The shell snippet and bullets now apply learnings only when the file exists and status is not COLLECTING, with a unified house-default fallback otherwise.
• ✅ Fixed: Learnings apply to all identities
  • The section is scoped to the personal/founder Typefully branch after identity resolution and explicitly excludes project upload-post drafts.

_{You can send follow-ups to the cloud agent here.}

^{Reviewed by Cursor Bugbot for commit 204ee94. Configure here.}

[sentry]

Bug: The script incorrectly treats the file path $PREFS_PATH as a directory when constructing the path for $LEARN, causing the feature to be silently disabled.
_{Severity: MEDIUM}

Suggested Fix

The path to learnings.md should be constructed relative to the directory containing the preferences.json file. Use dirname "$PREFS_PATH" to get the directory part of the path. For example: LEARN="$(dirname \"$PREFS_PATH\")/social-metrics/learnings.md".

Prompt for AI Agent

Review the code at the location below. A potential bug has been identified by an AI
agent. Verify if this is a real issue. If it is, propose a fix; if not, explain why it's
not valid.

Location: claude-ops/skills/ops-socials/SKILL.md#L121-L122

Potential issue: The bash snippet constructs the path
`LEARN="$PREFS_PATH/social-metrics/learnings.md"`. However, `$PREFS_PATH` is
consistently defined throughout the codebase as the full path to the `preferences.json`
file, not a directory. This results in an invalid path like
`.../preferences.json/social-metrics/learnings.md`. Consequently, the file existence
check `[ -f "$LEARN" ]` will always fail, and the `cat "$LEARN"` command will never
execute, silently disabling the "learnings" feature.

_{Did we get this right? 👍 / 👎 to inform future reviews.}

[auroracapital]

Heads up: this PR targets main. Per branch policy, please retarget to dev or confirm this is an intentional changelog/release PR.