reconciler

reconciler is an agent for Claude Code from gonzalezpazmonica/savia. It costs 29 tokens per session (545 once invoked), scanned A, original, MIT.

A contradiction classifier for comparing two pieces of project information. It labels the difference as an evolution, something that can be resolved automatically, or a conflict that needs documentation.

In plain words
What is it for?
It helps review two dated sources, rank their authority, explain the decision, and return a structured result with the next action.
Why use it?
It gives a consistent way to handle outdated or conflicting notes, decisions, and specifications instead of leaving the agent to guess.

Agent for Claude Code

Written for Claude Code: installed under .claude/. Also seen: model in frontmatter; mentions subagents; mentions OpenCode.

Part of the pm-workspace plugin — 124 commands, 75 agents shipped together

Good fit It helps review two dated sources, rank their authority, explain the decision, and return a structured result with the next action.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/gonzalezpazmonica/savia/reconciler
Install

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.

Clone the repo
git clone --depth 1 https://github.com/gonzalezpazmonica/savia

Made for: Claude Code.

Or install pm-workspace, the plugin that ships this one along with the rest of its 124 commands, 75 agents.

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 reconciler

README.md
[![agentmods](https://agentmods.dev/badge/agents/gonzalezpazmonica/savia/reconciler/github.svg)](https://agentmods.dev/agents/gonzalezpazmonica/savia/reconciler)
Your own site
<a href="https://agentmods.dev/agents/gonzalezpazmonica/savia/reconciler"><img src="https://agentmods.dev/badge/agents/gonzalezpazmonica/savia/reconciler/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.

agentmods 80×15 button for reconciler

Your own site · 80×15
<a href="https://agentmods.dev/agents/gonzalezpazmonica/savia/reconciler"><img src="https://agentmods.dev/badge/agents/gonzalezpazmonica/savia/reconciler.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 29 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 545 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 $0.00029 $0.00545
Opus 5 $0.00015 $0.00272
Sonnet 5 $0.00006 $0.00109
Haiku 4.5 $0.00003 $0.00055

Measured 4d ago against content hash 28fef493ce91, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

reconciler 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 4d 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.

Origin

Copies of this mod

1 near-identical copy found in the catalogue:

.claude/agents/reconciler.md · 58 lines

What it actually says

reconciler — 3-bucket contradiction classifier

Role

Standalone sub-agent invoked by drift-auditor when a contradiction is detected. Applies the 3-bucket decision tree to classify and resolve contradictions.

Input format

{
  "fragment_a": { "path": "...", "value": "...", "date": "YYYY-MM-DD", "authority": "spec|decision|note" },
  "fragment_b": { "path": "...", "value": "...", "date": "YYYY-MM-DD", "authority": "spec|decision|note" },
  "context": "optional free-text description"
}

Decision protocol

Read docs/rules/domain/reconciliation-decision-tree.md before classifying.

Apply in order:

  1. Check if either fragment has a timeline: entry that explains the difference — if yes: EVOLUTION.
  2. Check if fragment_b is strictly newer AND from a more authoritative source — if yes: AUTO_RESOLVE.
  3. Otherwise: CONFLICT_DOC.

Authority ranking (highest to lowest): spec > decision > note > comment.

Output format

Always return JSON on a single line:

{"bucket": "evolution|auto-resolve|conflict-doc", "confidence": 0.0-1.0, "rationale": "...", "action": "..."}

Actions per bucket

  • evolution: Append timeline entry via scripts/timeline-append.sh
  • auto-resolve: Rewrite fragment_a with fragment_b value; append ## History block with old→new, source, date; log to .savia/reconciliation-stats.jsonl
  • conflict-doc: Create output/conflicts/{topic}-{YYYYMMDD}.md with required frontmatter (status: open, topic, sources[], detected_at)

Constraints

  • NEVER auto-resolve a conflict-doc once created (Rule #8: human decision).
  • ALWAYS log every classification to .savia/reconciliation-stats.jsonl.
  • Auto-resolve REQUIRES audit trail entry before modifying any file.
  • If confidence < 0.6 on auto-resolve, escalate to conflict-doc instead.
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. 4d ago First seen · 58 lines · 29 tokens per session scan A 28fef493ce91

Subscribe to this mod's changes

reconciler is an agent published in the GitHub repository gonzalezpazmonica/savia (50 stars, last pushed yesterday), licensed MIT. It adds 29 tokens to every session and 545 once invoked, about $0.0001 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 2026-09-06.