fix: preserve content after frontmatter in parse_instinct_file()

[ericcai0814]

Summary

• parse_instinct_file() appended the instinct and reset state when frontmatter ended (---), before content lines could be collected — causing silent data loss of Action, Evidence, Examples sections during import.
• Fix: only set in_frontmatter = False at frontmatter close; defer append to next frontmatter start or EOF, where the logic already exists.
• Added pytest test covering multi-section instincts, single instinct, and empty content edge case.

Fixes #148

Reproduction

from parse_instinct_file import ...

content = """
---
id: test-1
trigger: "when coding"
confidence: 0.9
---

## Action
Always write tests first.

## Evidence
TDD improves quality.
"""

# Before fix: [{'id': 'test-1', 'content': ''}]
# After fix:  [{'id': 'test-1', 'content': '## Action\nAlways write tests first.\n\n## Evidence\nTDD improves quality.'}]

Test plan

• pytest skills/continuous-learning-v2/scripts/test_parse_instinct.py -v — 3 passed
• Manual verification with multi-instinct file — all content preserved
• Empty content edge case — no error

Summary by CodeRabbit

• Bug Fixes

  • Improved frontmatter parsing to correctly handle instinct boundary processing and content finalization

• Tests

  • Added comprehensive unit tests for instinct parsing covering multiple instincts, single instinct, and empty content scenarios to ensure parsing reliability

[coderabbitai]

📝 Walkthrough

Walkthrough

The changes fix a bug in the instinct file parser where content following frontmatter boundaries was lost during import. The parser now defers instinct finalization until the next frontmatter boundary or end-of-file, ensuring all content is preserved. Unit tests validate the fix across multiple scenarios.

Changes

Cohort / File(s)	Summary
Bug Fix skills/continuous-learning-v2/scripts/instinct-cli.py	Modified parse_instinct_file() to defer appending instincts until the next frontmatter boundary or EOF, rather than immediately when closing frontmatter is detected. This preserves content that follows the --- delimiter.
Tests skills/continuous-learning-v2/scripts/test_parse_instinct.py	Added comprehensive unit tests validating parse_instinct_file() correctly preserves content across multiple instincts, single instincts, and empty content scenarios.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~12 minutes

Poem

🐰 A parser's quest, oh what a tale,
Where frontmatter ends and content would pale,
But now with deference, 'til boundaries end,
Each instinct's wisdom flows without bend,
Tests verify this fix rings true and clear! ✨

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 25.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately describes the main fix: deferring content preservation in parse_instinct_file() to prevent loss of sections after frontmatter closure.
Linked Issues check	✅ Passed	The PR implementation directly addresses issue #148 by deferring in_frontmatter state change to preserve post-frontmatter content during instinct file imports.
Out of Scope Changes check	✅ Passed	All changes are scoped to fixing the parse_instinct_file() function and adding corresponding unit tests; no unrelated modifications detected.

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

✨ Finishing touches

• 📝 Generate docstrings

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment

---

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.}

[affaan-m]

Approved

Excellent fix. Verified the parsing logic end-to-end:

Fix correctness

The state machine flow is now:

1. --- (start frontmatter) → append previous instinct if exists, reset state, in_frontmatter = True
2. YAML key:value lines → populate current dict
3. --- (end frontmatter) → in_frontmatter = False (fix: no longer appends here)
4. Content lines → collect into content_lines
5. Next --- → back to step 1, which appends with collected content
6. EOF → append final instinct with collected content

The bug was step 3 duplicating step 1's append logic, but before step 4 could run. Removing it is the correct fix.

Test quality

• 3 focused test cases covering multi-instinct, single instinct, and empty content
• Smart use of importlib to handle the hyphenated filename
• Assertions check for specific content strings, not just non-empty
• Tests are self-contained with inline fixtures

PR quality

• Minimal diff (-5 lines of logic)
• Clear commit message with conventional format
• Links to issue (#148)
• Includes reproduction steps and before/after comparison

Well done @ericcai0814.

[maurez83630-cmyk]

github-recovery-codes.txt