codex-agent-mem

v1.0.x verification results

These are reproducible, sanitized results generated from synthetic fixtures.

Execution context:

• Runtime: Codex Desktop
• Model: not part of benchmark; synthetic fixtures only
• Reasoning effort: not part of benchmark; synthetic fixtures only

Snapshot

Scenario	Source tokens	Pack tokens	Saved	not_modified	Tools	Lazy init	Read-only
Small project continuity	1,841	253	86.26%	true	4	false->true	true
Medium agent workflow	4,855	270	94.44%	true	4	false->true	true
Large repeated audit	9,731	269	97.24%	true	4	false->true	true
Sub-agent handoff example	6,523	276	95.77%	true	4	false->true	true

Token savings by scenario

Small project continuity

A short project where the agent should remember objective, constraints and open work without repeating the whole discussion.

• Source context: ~1,841 tokens
• Memory pack: ~253 tokens
• Tokens not resent: ~1,588
• Estimated savings: 86.26%
• Pack hash: 700b30ab9323f77d7c8366ebe09bcb7084645bbc31dcf7092fae5a15ca568b48

source [############################] 100.00% pack [####……………………] 13.74% saved [########################….] 86.26%

Medium agent workflow

A realistic multi-step implementation with repeated decisions, pending work and DoD requirements.

• Source context: ~4,855 tokens
• Memory pack: ~270 tokens
• Tokens not resent: ~4,585
• Estimated savings: 94.44%
• Pack hash: b5c6ecd9ec15c6463cbfd924247bba500dca7bbb3834144afc6f7a7051d00aac

source [############################] 100.00% pack [##……………………..] 5.56% saved [##########################..] 94.44%

Large repeated audit

A long audit where the same constraints and decisions would normally be re-sent many times.

• Source context: ~9,731 tokens
• Memory pack: ~269 tokens
• Tokens not resent: ~9,462
• Estimated savings: 97.24%
• Pack hash: 6cc762e877b54779ac835bf0820712bbd3e4971472926a3cba110557221e6664

source [############################] 100.00% pack [#………………………] 2.76% saved [###########################.] 97.24%

Sub-agent handoff example

A project where an explorer sub-agent audits context and a worker sub-agent implements a bounded change.

• Source context: ~6,523 tokens
• Memory pack: ~276 tokens
• Tokens not resent: ~6,247
• Estimated savings: 95.77%
• Pack hash: 1e445aab0917fb3baef42a14ad8443e4f94b5758b84601fde013f07ae7b879e8

source [############################] 100.00% pack [#………………………] 4.23% saved [###########################.] 95.77%

Sub-agent example:

• explorer: Audit the memory surface and identify which facts must never be dropped.
• worker: Implement the bounded runtime-efficiency patch without touching documentation.

Repeated context avoided

known_pack_hash lets the agent ask whether a pack changed before re-sending it.

Scenario	Result
Small project continuity	not_modified=true
Medium agent workflow	not_modified=true
Large repeated audit	not_modified=true
Sub-agent handoff example	not_modified=true

Runtime safety

Metric	Result
Minimal profile tools	mem_open_work, mem_completion_check, mem_context_pack, mem_health_runtime
Tool count in minimal profile	4
Lazy initialization before DB-backed tool	false
Lazy initialization after context pack	true
Mutating tool tested in read-only mode	mem_snapshot_create
Mutating tool blocked	true

This is the v1.0.0 fixture baseline. In v1.0.1, the minimal profile also includes mem_session_list, mem_scope_resolve, and mem_bootstrap_context so broad workspaces can resolve scope before loading active context.

Response diet

Text shown to the model can be kept compact while the structured payload remains available to MCP clients.

Scenario	Compact text chars	Balanced text chars	Verbose text chars
Small project continuity	160	209	20,715
Medium agent workflow	160	209	42,943
Large repeated audit	160	209	76,114
Sub-agent handoff example	160	209	54,625

Telemetry smoke

• Telemetry mode tested: summary
• Events captured: process_start, initialize, tools_list, tool_call, stdin_eof, process_exit
• Stores memory content: false

Interpretation

These numbers are not a universal guarantee. They show reproducible behavior on public synthetic fixtures. The expected value is highest when an agent would otherwise resend repeated project context across sessions.

This site is open source. Improve this page.