Borrowing it
Nothing to install: this file belongs to marciopuga/cog. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/marciopuga/cog/main/.claude/commands/reflect.mdgit clone --depth 1 https://github.com/marciopuga/cogWrote 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/commands/marciopuga/cog/reflect)<a href="https://agentmods.dev/commands/marciopuga/cog/reflect"><img src="https://agentmods.dev/badge/commands/marciopuga/cog/reflect/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/commands/marciopuga/cog/reflect"><img src="https://agentmods.dev/badge/commands/marciopuga/cog/reflect.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 | $0.00047 | $0.02115 |
| Opus 5.5 | $0.00019 | $0.00846 |
| Sonnet 5.5 | $0.00009 | $0.00423 |
| Haiku 4.5 | $0.00005 | $0.00212 |
Grade A, and why
reflect 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 2mo 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 — 167 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Cog Reflect
Self-reflection and memory consolidation. Past-facing — reads memory, fixes contradictions, distills patterns. Reflect never sees transcripts: everything it knows comes from what was written to memory, so an interaction that was never captured is invisible here.
Take your time. This is a deep session. Read broadly, cross-reference, and ACT on findings. You're the maintainer of the knowledge base.
Memory Path
All files under the resolved memory path: $COG_HOME/memory/ if COG_HOME is set, otherwise ~/cog/memory/.
Orientation (run first)
Scope your work before reading files:
- Find files changed since the last run — read
memory/cog-meta/run-log.mdfor the last/reflectentry and scope to files modified since that date (no entry → default to the last 7 days) - Read
memory/domains.yml, thenINDEX.mdonly for domains that have changed files — the index gives L0 + line counts per file. Don't grep L0 headers across the whole tree. Never edit anINDEX.md— housekeeping regenerates them; a week of drift in line counts is harmless. - Check observation entry counts (archival threshold = 50)
Focus on recently-changed files. Skip unchanged ones.
Minimum Data Check
Before proceeding, verify there's enough material to work with:
- If total observations across all domains < 5: stop. Say "Not enough data yet. Keep capturing observations and run again when you have more material."
- If no files were modified in the last 7 days: flag it. "Memory hasn't been updated recently. Consider capturing some observations first."
- If patterns.md is empty and observations < 10: say "Too early to consolidate. You need ~10+ observations before patterns emerge."
Don't produce low-quality output from insufficient data. It's better to say "not yet" than to force weak patterns.
Files to Read
memory/cog-meta/self-observations.mdmemory/cog-meta/patterns.mdmemory/cog-meta/action-items.md(open[housekeeping]routes andpri:lowideas)memory/cog-meta/scenarios/*.md(active scenarios, for the retrospective step)- All domain
observations.mdfiles - All domain
action-items.mdfiles - All
hot-memory.mdfiles
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.
- 2mo ago Changed · -1 lines · +4 tokens per session 80e8f8b6e419
- 2mo ago First seen · 168 lines · 43 tokens per session scan A 79c6de17c7ef
reflect is a command published in the GitHub repository marciopuga/cog (377 stars, last pushed 2mo ago), licensed MIT. It adds 47 tokens to every session and 2,115 once invoked, about $0.0002 per session on Opus 5.5. 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.
reflect compared with commands from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| reflectmarciopuga/cog | Command | 377 | 2mo ago | MIT | A | 47 | Claude Code |
| minutes-ideassilverstein/minutes | Command | 1.6k | yesterday | MIT | A | 48 | Claude Code, OpenCode |
| learnReflexioAI/claude-smart | Command | 791 | 3d ago | Apache-2.0 | A | 14 | Claude Code |
| memory-storeverygoodplugins/mcp-automem | Command | 66 | today | MIT | A | 10 | Claude Code |
| analyze-contextRohiRIK/OpenLtm | Command | 29 | 8d ago | MIT | A | 39 | Claude Code |
| cc-memoryTheLobbi/claude | Command | 21 | 7d ago | MIT | A | 0 | Claude Code |
| lians-recallLians-ai/Lians | Command | 17 | 4d ago | Apache-2.0 | A | 19 | Claude Code |