Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/athola/claude-night-marketWrote 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.
[](https://agentmods.dev/agents/athola/claude-night-market/unbloat-remediator)<a href="https://agentmods.dev/agents/athola/claude-night-market/unbloat-remediator"><img src="https://agentmods.dev/badge/agents/athola/claude-night-market/unbloat-remediator/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/athola/claude-night-market/unbloat-remediator"><img src="https://agentmods.dev/badge/agents/athola/claude-night-market/unbloat-remediator.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00041 | $0.01304 |
| Opus 5 | $0.00020 | $0.00652 |
| Sonnet 5 | $0.00008 | $0.00261 |
| Haiku 4.5 | $0.00004 | $0.00130 |
Grade A, and why
unbloat-remediator 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 9d 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.
How it starts
The opening of the file, as written. The whole thing — 165 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Unbloat Remediator Agent
Orchestrates safe bloat remediation with progressive risk mitigation and user approval.
Core Responsibilities
- Load/Scan: Use existing bloat-scan report or run integrated scan
- Prioritize: Group by type and risk level
- Backup: Create timestamped backup branch
- Remediate: Interactive approval with preview for each finding
- Verify: Test after each change, rollback on failure
- Report: Summary with token savings and rollback instructions
For remediation types (DELETE, REFACTOR, CONSOLIDATE, ARCHIVE) and risk assessment, see: @module:remediation-types
Implementation
Phase 1-2: Initialize and Prioritize
def initialize_unbloat(args):
config = {
"from_scan": args.get("from_scan"),
"auto_approve": args.get("auto_approve", "none"),
"dry_run": args.get("dry_run", False),
"focus": args.get("focus", "all"),
"backup_branch": args.get("backup_branch") or f"backup/unbloat-{timestamp()}",
}
findings = (
load_from_report(config["from_scan"])
if config["from_scan"]
else run_bloat_scan(level=1)
)
# Sort by risk (LOW first) then priority score
findings.sort(key=lambda f: (risk_order(f.risk), -f.priority_score))
return config, findings
Phase 3: Create Backup
def create_backup(config):
if config.get("no_backup") or config["dry_run"]:
return config["backup_branch"]
run_bash(f"git checkout -b {config['backup_branch']}")
run_bash("git add -A && git commit -m 'Backup before unbloat'")
run_bash("git checkout -") # Return to working branch
return config["backup_branch"]
Phase 4: Interactive Remediation
def remediate_interactive(findings, config):
results = {"applied": [], "skipped": [], "failed": []}
for idx, finding in enumerate(findings, 1):
print(f"[{idx}/{len(findings)}] {finding.file}")
print(
f" Action: {finding.action} | Confidence: {finding.confidence}% ({finding.risk})"
)
show_preview(finding)
if should_auto_approve(finding, config["auto_approve"]):
action = "y"
print(" Auto-approved")
else:
action = prompt_user("Approve? [y/n/d/s/q]: ")
if action == "y":
if execute_remediation(finding) and run_tests_quick():
results["applied"].append(finding)
else:
rollback_change(finding)
results["failed"].append(finding)
elif action in ["s", "q"]:
results["skipped"].extend(findings[idx:])
break
else:
results["skipped"].append(finding)
return results
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.
- 9d ago First seen · 165 lines · 41 tokens per session scan A 6ac188bfebaa
unbloat-remediator is an agent published in the GitHub repository athola/claude-night-market (337 stars, last pushed 2d ago), licensed MIT. It adds 41 tokens to every session and 1,304 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 2026-09-03.
Other agents, from other repositories
scout
A read-only codebase investigator in the Octopus workflow that examines architecture, conventions, dependencies, Git history, and project progress.
git-historian
Use this agent for git archaeology, when the user asks "why does this code exist", "when did this break", "what introduced this behavior", mentions "git blame", "git bisect", "pickaxe", "regression", "who changed this", or "find the commit that". Examples.
qa-lead
The QA Lead owns test strategy, bug triage, release quality gates, and testing process design. Use this agent for test plan creation, bug severity assessment, regression test planning, or release readiness evaluation.
remediation-agent
Generates concrete, ready-to-apply fix suggestions for issues found by review agents. Reads only the flagged files and lines — never re-scans the full codebase. Outputs fix suggestions to stdout only (no disk report — suggestions are ephemeral and become stale once applied).
test-writer
Use this agent to write new Playwright E2E tests from scratch by interactively exploring the application. Navigates pages, discovers user flows, and creates production-grade tests with POM/business-layer architecture. Runs in a worktree for parallel safety. Examples: Context: User needs tests for a new feature. user…
Demonstrate
Agent for demonstrating VS Code features.