Borrowing it
Nothing to install: this file belongs to AGI4Sci/SciForge. 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/AGI4Sci/SciForge/main/.codex/skills/research-memory/SKILL.mdgit clone --depth 1 https://github.com/AGI4Sci/SciForgeWrote 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/skills/agi4sci/sciforge/research-memory)<a href="https://agentmods.dev/skills/agi4sci/sciforge/research-memory"><img src="https://agentmods.dev/badge/skills/agi4sci/sciforge/research-memory/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/skills/agi4sci/sciforge/research-memory"><img src="https://agentmods.dev/badge/skills/agi4sci/sciforge/research-memory.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 | $0.00036 | $0.01219 |
| Opus 5.5 | $0.00014 | $0.00488 |
| Sonnet 5.5 | $0.00007 | $0.00244 |
| Haiku 4.5 | $0.00004 | $0.00122 |
Grade A, and why
research-memory 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 22d 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 — 145 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Coordinate research memory updates through this skill, GitHub conventions, and repository templates.
Use this skill when the user asks to record, update, summarize, or synchronize:
- Experiment outcomes, metrics, artifacts, or validation state
- Blockers, risks, failed runs, open questions, or next actions
- GitHub issue or PR summaries for research work
- Collaborator feedback that should be preserved as research evidence
- Status updates derived from confirmed project memory
Do not use this skill for ordinary code edits, local-only notes, or GitHub actions that do not involve research memory.
Core Rule
The agent MUST follow this order for every Research Memory update:
- Draft the proposed memory update, GitHub summary, status update, or artifact record.
- Ask the user to confirm the draft.
- Only after explicit confirmation, apply the file change or post/synchronize it to GitHub.
If the user asks to "just sync it" without seeing a draft, still produce a draft first and request confirmation.
No MCP
Research Memory is intentionally a skill plus conventions now. Do not call, configure, install, or recreate a project-memory MCP server or worker. Use normal file edits and GitHub issue/PR/comment workflows after confirmation.
Templates
Prefer the repository templates in docs/templates/research-memory/:
artifact-record.mdgithub-issue.mdgithub-pr.mdstatus-update.md
Fill only fields supported by the user's confirmed facts. Leave unknown fields as TBD or ask a short follow-up question when the field is required for GitHub output.
Hard Limits
The agent MUST NOT use GitHub or any connected workflow to:
- Automatically merge a pull request
- Close a major issue without explicit user approval
- Mark any result, experiment, or artifact as
validated - Publish a public claim about research results
When a user asks for one of these actions, draft the proposed wording or action and ask for explicit confirmation. Treat validated and public claims as high-risk labels requiring clear user authorization and supporting evidence.
GitHub Requirements
All GitHub-facing summaries created from Research Memory MUST include:
- Artifact ID
- Evidence level
Use concise labels such as:
Artifact ID: EXP-2026-06-25-rerank-a17Evidence level: preliminary
If either value is missing, ask for it before posting or preparing a PR.
GitHub PRs carry review. Use PR comments, PR descriptions, or review threads for critique, questions, and reviewer decisions. Do not treat a status page as the review channel.
Evidence Levels
Prefer one of these evidence levels unless the project defines a stricter vocabulary:
observation: seen in a run, log, issue, or collaborator reportpreliminary: supported by an artifact but not independently checkedreproduced: repeated with matching outcomevalidated: explicitly approved by the user as validated
Never infer validated from passing tests, a successful run, or a collaborator comment.
Workflow
- Identify the research event and whether GitHub output is requested.
- Gather required fields:
- Artifact ID
- Evidence level
- Summary
- Source evidence, such as logs, PR, issue, run ID, file path, or collaborator comment
- Requested target, such as local artifact record, GitHub issue, GitHub PR, PR comment, or status update
- Draft the exact update using the matching template.
- Ask the user to confirm.
- After confirmation, apply local file edits or post/synchronize to GitHub.
- Report what changed and where.
Minimum Examples
Experiment completed:
Draft memory update:
Artifact ID: EXP-2026-06-25-rerank-a17
Evidence level: preliminary
Summary: Reranker A17 completed on the held-out query set. It improved top-3 recall from 0.71 to 0.76, with latency increasing from 84 ms to 109 ms.
Evidence: run run-4821, metrics artifact metrics/rerank-a17.json
Target: Local artifact record and GitHub PR summary
Please confirm before I update files or prepare GitHub output.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 22d ago First seen · 145 lines · 36 tokens per session scan A 560d68efca1a
research-memory is a skill published in the GitHub repository AGI4Sci/SciForge (83 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 36 tokens to every session and 1,219 once invoked, about $0.0001 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 2026-10-02.
Other skills, from other repositories
forget
Remove a personal fact, preference, relationship detail, topic, or prior event from Muse's active memory and stop existing copies or automations from bringing it back. Use for explicit requests such as 'forget that', 'don't remember this about me', or 'remove that from your memory'. Do not use when 'forget it' merely…
mem0-oss-to-platform
Plan, then execute, a migration of a project from the mem0 open-source / self-hosted SDK (local Memory class) to the mem0 Platform / hosted SDK (MemoryClient). TRIGGER when: the developer wants to move, switch, or migrate mem0 off OSS/self-hosted to the hosted API, e.g. "migrate my mem0 setup to the platform", "switch…
Cortex
Operate Cortex, the LifeOS memory system — the typed Knowledge Archive (People, Companies, Ideas, Research with typed related: links) plus recall of prior work sessions, ISAs, and conversations. Search, add, harvest, develop, ingest, distill, graph-navigate, recall. USE WHEN cortex, knowledge, knowledge base, search…
memory
Use when the user asks to remember, recall, forget, update, search, or inspect durable OpenSquilla memory, including profile facts in USER.md and long-term notes in MEMORY.md or memory//.md.
ha-data-stores
Map of Hope Agent's local data stores and safe read-only query workflow. Use when the user asks where Hope Agent stores data, wants to inspect sessions/messages/memory/logs/background jobs/knowledge indexes/settings, asks the model to query local app data, or debugging requires checking persisted state. Trigger…
shellm
Reference for the shellm system — recursive LLM shell, identity management, memory, skills, trajectory, and all CLI tools. Use when working on shellm itself, debugging agent behavior, or understanding how the pieces fit together.