> Source URL: /demos/cs1-path/templates/template.feedback-harness
---
student_id:
project_id:
submission_date:
reviewer: # instructor or "agent-assisted"
---

# Feedback Harness: [STUDENT_ID] - [PROJECT_NAME]

## Purpose

This feedback harness is a **reviewer artifact** that bundles all context needed for AI and human reviewers to evaluate student work. It contains:

- Student profile and persona information
- Project specification and rubric
- Student's submitted code as an artifact
- Commit history and development process
- Agent instructions for generating personalized feedback

**Note**: This harness is typically created via a custom GitHub Classroom integration (not provided in this example repository). The integration automatically bundles student submissions, commit history, and links to relevant curriculum materials.

## Student-Facing Feedback

The feedback generated in this harness is extracted and delivered to the student as a separate file:

**Student-facing feedback**: [STUDENT_ID]-[PROJECT_ID].feedback.md (in `curriculum/students/` folder)

The student-facing file contains only the feedback content without internal context (rubric scores, commit history, agent instructions, etc.).

## Student Profile

See: [student persona](../students/student-[ID].persona.md)

**Quick summary:**

- Prior experience: [level]
- Strengths: [brief]
- Growth areas: [brief]

## Project Specification

See: [project spec](../projects/project-[NUMBER].spec.md)

## Rubric

See: [rubric](../projects/project-[NUMBER].rubric.md)

| Criterion     | Points Possible | Score | Notes |
| ------------- | --------------- | ----- | ----- |
| [Criterion 1] | X               |       |       |
| [Criterion 2] | X               |       |       |
| [Criterion 3] | X               |       |       |
| **Total**     | **XX**          |       |       |

## Commit History

Recent commits from student's repository:

- `[hash]`: [message] ([date])
- `[hash]`: [message] ([date])
- `[hash]`: [message] ([date])

## Code Snapshot

### [filename.py]

```python
# Student's submitted code
# Paste the relevant file(s) here
```

### [another_file.py] (if applicable)

```python
# Additional files
```

## Agent Instructions

When generating feedback:

1. **Be specific**: Reference line numbers and exact code
2. **Be constructive**: Focus on improvement, not just problems
3. **Acknowledge strengths**: Note what works well
4. **Connect to learning**: Link feedback to course outcomes
5. **Match the student**: Consider their persona and experience level
6. **Prioritize**: Focus on 2-3 key improvements, not every issue

### Feedback Format

Structure feedback as:

1. **Summary**: Overall assessment (2-3 sentences)
2. **What works well**: Specific positives
3. **Key improvements**: Prioritized suggestions with examples
4. **Next steps**: Concrete actions for the student

## Generated Feedback

**Note**: The feedback below is extracted and delivered to the student as `curriculum/students/[STUDENT_ID]-[PROJECT_ID].feedback.md`. The student-facing version omits internal context (rubric scores, commit history, agent instructions) and focuses solely on actionable feedback.

[Agent or instructor writes feedback below this line]

---

### Summary

[Overall assessment]

### What Works Well

- [Positive 1]
- [Positive 2]

### Key Improvements

#### 1. [Issue Title]

**Location**: Line X

**Current code**:

```python
# problematic code
```

**Suggestion**:

```python
# improved approach
```

**Why**: [Explanation connecting to learning outcome]

#### 2. [Issue Title]

[Same format]

### Next Steps

1. [Specific action]
2. [Specific action]


---

## Backlinks

The following sources link to this document:

- [Feedback Harness Template](/demos/cs1-path/index.path.llm.md)