SEPCC: Agent for Claude Code

.claude/agents/fcc-agent-fixer.md

fcc-agent-fixer is an agent for Claude Code from sepehrbayat/SEPCC. It costs 35 tokens per session (953 once invoked), scanned A, original, MIT.

An automated code-fixing agent that reads a DebugReport from fcc-agent-debugger and applies its findings in priority order. It works in a separate Git worktree and runs tests after each fix.

In plain words
What is it for?
Use it to address reported correctness, completeness, and process findings, verify them with pytest, and record failures when a fix cannot be safely applied.
Why use it?
It keeps fixes isolated from the main working directory and checks whether each change still passes the test suite. Failed fixes are skipped and documented rather than left partially applied.

Agent for Claude Code

Written for Claude Code: installed under .claude/. Also seen: model in frontmatter.

This is sepehrbayat/SEPCC's own configuration. It tells Claude Code how to work on SEPCC itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything SEPCC configures →

Reuse

Borrowing it

Nothing to install: this file belongs to sepehrbayat/SEPCC. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/sepehrbayat/SEPCC/main/.claude/agents/fcc-agent-fixer.md
Clone the repo
git clone --depth 1 https://github.com/sepehrbayat/SEPCC

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for fcc-agent-fixer

README.md
[![agentmods](https://agentmods.dev/badge/agents/sepehrbayat/sepcc/fcc-agent-fixer.svg)](https://agentmods.dev/agents/sepehrbayat/sepcc/fcc-agent-fixer)
Your own site
<a href="https://agentmods.dev/agents/sepehrbayat/sepcc/fcc-agent-fixer"><img src="https://agentmods.dev/badge/agents/sepehrbayat/sepcc/fcc-agent-fixer.svg" alt="Measured on agentmods" height="20"></a>
Per session 35 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 953 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00035 $0.00953
Opus 5 $0.00017 $0.00477
Sonnet 5 $0.00007 $0.00191
Haiku 4.5 $0.00003 $0.00095

Measured 8d ago against content hash 1eb8ce4e56c6, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-07, from the pricing page.

Security

Grade A, and why

fcc-agent-fixer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/agents/fcc-agent-fixer.md · 132 lines

How it starts

The opening of the file, as written. The whole thing — 132 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are the FCC Agent Fixer. Your job is to read a DebugReport produced by fcc-agent-debugger and apply fixes in priority order, verifying each fix with tests.

Setup: Isolated Worktree

  1. Read the DebugReport from .fcc/debugger/reports/<task-slug>.json.

  2. Create an isolated git worktree:

    git worktree add --detach .claude/worktrees/fixer-<task-slug> HEAD
    cd .claude/worktrees/fixer-<task-slug>
    
  3. If git worktree add fails for any reason:

    • Write a file .fcc/debugger/fixes/<task-slug>-FAILED.md describing the failure.
    • Do NOT modify any files in the main working tree.
    • Exit immediately.

Fix Protocol

Process findings in fix_priority order (already sorted by severity: correctness findings first, then completeness, then process).

For each finding:

  1. Read the file referenced in the finding's file field.

  2. Apply the fix described in suggested_fix (or devise your own fix if the suggestion is insufficient).

  3. Immediately run the test suite and verify it passes:

    uv run pytest
    
  4. If tests fail after the fix: revert the change and mark the finding as fixes_skipped with a reason explaining the test failure. Do NOT proceed to the next finding until tests pass again (either the fix worked, or it was reverted).

  5. If no test suite exists: apply the fix but set verified: false in the fix entry.

After all fixes are applied:

Run the full test suite one final time to confirm everything passes:

uv run pytest

Commit

Commit all fixes in a single commit:

git commit -m "fix: resolve [N] issues from task '[task_name]' [debugger]"

The commit body must list each fix ID and the action taken:

CORR-001: Added null check for user_id before database query
COMP-001: Implemented email validation, removed TODO marker
PROC-001: Added type annotations to validate_token

Output

Write a FixReport as JSON to .fcc/debugger/fixes/<task-slug>.json.

Read the full file on GitHub · 132 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 8d ago First seen · 132 lines · 35 tokens per session scan A 1eb8ce4e56c6

Subscribe to this mod's changes

fcc-agent-fixer is an agent published in the GitHub repository sepehrbayat/SEPCC (38 stars, last pushed 16d ago), licensed MIT. It adds 35 tokens to every session and 953 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other agents, from other repositories

architect

Deep reasoning agent. Architecture decisions, security reviews, complex debugging, performance analysis, system design, race conditions, data modelling. Use when getting it wrong has high cost.

FerroxLabs/ijfw · 36 tokens

oracle

Deep code analyst and debugger. Use for root cause analysis, understanding complex code, reading and interpreting diffs, tracing bugs, and deep comprehension tasks. Avoid for simple lookups or first-attempt fixes.

znlgis/my-opencode-deepseek-config · 43 tokens

debugger

Use when encountering any bug, test failure, unexpected behavior, or error - before proposing fixes.

herbert-julio-azion/specialist-agent · 22 tokens

code-reviewer-bug

name: code-reviewer-bug description: Specialized code reviewer for bug patterns — null safety, race conditions, resource leaks, logic and error-handling defects. Returns scored findings (severity × impact × confidence). skills: code-review model: inherit.

zxpmail/ReqForge · 0 tokens

debugger

Systematic debugger using the Iron Law: no fix without confirmed root cause. Reproduces errors, traces execution paths, forms and verifies hypotheses, then implements…

KevinZai/commander · 35 tokens

feedback-observer

name: feedback-observer description: Dispatched after failures (compile error, review fail, verification fail) OR user corrections/feedback. Auto-scores Skill dimensions on failure. Uses feedback-writer skill to record feedback. skills: feedback-writer model: opus color: blue.

zxpmail/ReqForge · 0 tokens