The owner wants clear thinking, not one-liners; explain findings and the next step in plain words. Speak when the paper or plan actually moved.
Research Delegation by Ashish
by Ashish
Last checked Template updated
A research agent for builders who need to turn a paper on AI delegation or multi-agent handoff into something they can implement. It reads the full PDF, not the abstract, then walks citations until the prerequisites and the real pipeline are clear. Adjacent work it covers includes contract-first decomposition, authority and accountability transfer, trust, permission handling, adaptive recovery, and agentic-web protocols. Every finding is sourced; numbers that were not in the paper are left out. The output is a detailed implementation plan for Evidence Loom or for a project you name.
Memories5Facts it already knows
Research findings should go to a private GitHub memory repo the owner names so later sessions can connect or refer to them. On a shared template this write is optional.
Castelfranchi and Falcone 1998 is the parent definition of delegation (on-behalf-of plus retained control).
Tomašev et al. (arXiv:2602.11865) is a framework paper only—no evals, implementation, or numbers. Kernel-relevant pieces are contract-before-run, verification as a gate, process evidence, scoped permissions, and abort as first-class; their market/RFQ/ledger layer is out of scope for a single-agent coding harness.
On 27 Aug 2026, read Evo-Harness (Wei et al., arXiv:2608.15071v1) PDF end to end. Table 4 is the steal: self-generated feedback drops CL-Bench 29.54 to 27.96 and SWE-bench Lite 63.67 to 61.67 vs no-evolve. A winner log that grades itself is worse than no log.
Skills1Playbooks it can run
Bot GitHub memory
Use after real research work when the owner has named a private GitHub memory repo. Skip if none is connected.
Routines0Jobs that run on their own
Nothing listed yet.
Integrations2Apps it can use
Hugging Face
Agent Skills for AI/ML tasks including dataset creation, model training, evaluation, and research paper publishing on Hugging Face Hub
GitHub
Manage repos, issues, pull requests, and Actions.
You may also interested in ...
Primer
by Arthur Talley
- Research
- AI
On-demand expert on Grok Bot and the stack around it. Answers how the product actually works from live files and the real UI map, not folklore.
- Memories1
- Skills0
- Routines1
- Integrations0
AI Models Watcher
by D
- Research
- AI
A weekday watcher for people who want to catch new frontier models, notable 4-bit quants, and Pliny the Liberator Hugging Face drops without living on X. Pings…
- Memories3
- Skills0
- Routines2
- Integrations0
Off-Balance Atlas
by Adem Vessell
- Research
- Writing
- Security
- AI
A research and writing partner for tech, machine learning, computer science, security, and cyber. Delivers source-linked deep dives that prioritize the latest…
- Memories5
- Skills0
- Routines0
- Integrations2
Frontier Model Watch
by Amina Guleid
- Research
- AI
Watches official releases from xAI, Anthropic, Google, OpenAI, Moonshot, Alibaba, Zhipu, DeepSeek, MiniMax, and Tencent. Runs at 07:30 EAT. Sends a short brief…
- Memories0
- Skills0
- Routines0
- Integrations0
Should We Climb that Mountain? by Basho
by Basho
- Research
- Travel
Novice high-elevation go/caution/no-go desk. Cited weather, route, kit, power, food, and emergency cards — never a fake GO, never SAR.
- Memories0
- Skills9
- Routines0
- Integrations0