agentleFS
search-atlas-groupsearch-atlas-group/amm-founding-circle/skills/instinct-bridge/SKILL.md

instinct-bridge

Skill0 starsChanged 5 months agoMITView on GitHub

Bridge GSD milestone patterns/lessons into the instinct pipeline as candidates

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 7 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read a GSD LEARNINGS.md or SUMMARY.md artifact, extract all Patterns and Lessons sections, and feed them through the instinct deduplication …

    line 13
  2. Distinctive:

    If no file path is provided, exit with:

    line 27
  3. Distinctive:

    Read the file and confirm it exists and contains at least one of these headings:

    line 35
  4. Distinctive:

    If none of these sections exist, report:

    line 43
  5. Distinctive:

    If N=0 for any section, omit that section from the report (don't print an empty "### New Instincts Created (0)").

    line 94
  6. Distinctive:

    Never fabricate patterns. Only process what is in the file.

    line 98
  7. Distinctive:

    Do not promote candidates automatically — they enter the pipeline at status: candidate with confidence 0.45.

    line 99
  8. Very specific:

    If the bridge library is not importable, report the import error clearly and suggest running: ls ~/.claude/instincts/lib/

    line 101

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md
<objective> Read a GSD LEARNINGS.md or SUMMARY.md artifact, extract all Patterns and Lessons sections, and feed them through the instinct deduplication + creation pipeline. Each non-duplicate, non-trivial bullet becomes a candidate instinct file in the appropriate scope directory. </objective>

Instructions

Step 1: Parse arguments

Extract the summary file path and optional project slug from the invocation arguments:

<summary-file-path>         — required; path to GSD LEARNINGS.md or SUMMARY.md
--project <slug>            — optional; overrides auto-derived project slug

If no file path is provided, exit with:

Error: provide a path to a GSD LEARNINGS.md or SUMMARY.md file.
Usage: /instinct-bridge <summary-file-path> [--project <slug>]

Step 2: Validate the file

Read the file and confirm it exists and contains at least one of these headings:

  • ## Patterns
  • ## Lessons Learned
  • ## Lessons
  • ## What Worked
  • ## What Didn't Work
  • ## Surprises

If none of these sections exist, report:

No extractable sections found in <filepath>.
Expected: ## Patterns, ## Lessons Learned, ## What Worked, ## What Didn't Work, or ## Surprises

Step 3: Run the bridge

python3 -c "
import sys
sys.path.insert(0, '$HOME/.claude/instincts/lib')
from gsd_bridge import bridge_milestone_learnings
import json

result = bridge_milestone_learnings(
    '$SUMMARY_FILE',
    '$HOME/.claude/instincts/global/',
    $PROJECT_DIR_ARG
)
print(json.dumps(result, indent=2))
"

Where:

  • $SUMMARY_FILE is the provided file path (shell-escaped)
  • $PROJECT_DIR_ARG is either None or '$HOME/.claude/instincts/projects/<slug>/' if --project was given

Step 4: Produce the report

Format the results as:

## GSD Bridge Report

**Source**: <filepath>
**Project**: <slug or "global">

### New Instincts Created (N)
- INS-XXX | "<name>" | confidence: 0.45 | from: "<pattern text snippet>"

### Merged with Existing (N)
- INS-XXX | "<existing_name>" | similarity: 0.XX | pattern: "<snippet>"

### Skipped (N)
- "<pattern text>" — reason: <too vague | duplicate | too short>

---
Next: run /instinct-learn-eval to score and potentially promote these candidates.

If N=0 for any section, omit that section from the report (don't print an empty "### New Instincts Created (0)").

Important Constraints

  • Never fabricate patterns. Only process what is in the file.
  • Do not promote candidates automatically — they enter the pipeline at status: candidate with confidence 0.45.
  • The global instincts dir is always searched for duplicates, even when writing to a project dir.
  • If the bridge library is not importable, report the import error clearly and suggest running: ls ~/.claude/instincts/lib/
  • mkdir -p is handled by the library — no manual directory creation needed.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.