BotCurrent↗
← Back to examples

User report · Research across apps

Maintain a searchable wiki from work activity

Source-backed example. Practical decision guidance is proposed and untested; this is not a full implementation guide.

The problem

Useful work knowledge disappears into separate conversations, code changes and documents, making prior context difficult to recover.

Agent job

Ingest scoped work sources on schedule into a personal wiki, entity graph and retrieval index.

Expected output

A searchable personal work wiki and index assembled from scoped sources.

Where this framing comes from
  • Core problem: Editorial purpose framing
  • Agent job: Editorial purpose framing
  • Expected output: Editorial purpose framing
  • Relevant job roles: Editorial purpose framing
  • Benefit type: Editorial purpose framing
Editorial framing and outcome limits

Describes the task’s purpose and intended output; measured benefit, safe unsupervised operation and successful delivery are not established by this framing.

Editorial synthesis based on the linked source. This framing does not add evidence of successful delivery or results.

Case-specific decision guidance · untested

Case-specific decision guidance and first checks are original proposals. The linked source supports the attributed task; this added guidance has not been executed.

Read the attributed source evidence and original caveats below ↓

Why use this

  1. A project decision can be spread across a conversation, a code change and an evolving document. When someone revisits the issue, a confident summary may hide that the discussion was tentative or later superseded.

  2. A private knowledge packet could link decisions to dated evidence and expose conflicting statements. Reviewers could recover the relevant context without treating a generated wiki entry as a new authority or widening access to the underlying work.

Why an agent / simpler alternative

Useful judgment

Connect differently named project entities and distinguish proposed decisions from confirmed changes.

Simpler alternative

Use a manually maintained decision log when a team has few sources and clear ownership.

Unsuitable when

Source permissions cannot be preserved or colleagues' restricted content would enter a broader index.

Proposed inputs

  1. Name

    Work-source sample

    Owner

    Authorized source owners

    Minimum scope

    Approved excerpts with timestamps, source IDs and access labels

    Quality check

    Remove excluded channels, repositories and documents.

  2. Name

    Entity glossary

    Owner

    Project lead

    Minimum scope

    Project names, aliases and known decision owners

    Quality check

    Label uncertain identity matches rather than merging automatically.

Access and setup — proposed or unknown

Proposed minimum
  1. Read-only exported excerpts in a private staging index with explicit retention boundaries.

Compatibility unknowns
  1. Unknown: current runtime, connector permission inheritance and deletion propagation into indexes.

External actions requiring approval
  1. Expanding ingestion scope, persistent access, sharing wiki pages and changing source permissions.

Expected deliverable and review checks

Proposed artifact

Proposed wiki entries with claim-level citations, chronology, access labels and unresolved contradictions.

Critical checks
  1. A reader can trace each decision to an authorized source.

  2. Superseded statements remain distinguishable from current proposals.

Missing output evidence

No new workflow run or output inspection is performed for this editorial layer. The linked implementation evidence retains the original source-specific outcome and artifact limits.

Failure signals and proposed responses

  1. Signal

    Two people or projects with similar names are merged.

    Why it matters

    The wiki can attribute decisions to the wrong owner.

    Proposed response

    Split the entity and request a source-backed identity match.

  2. Signal

    A restricted excerpt appears in an unrestricted summary.

    Why it matters

    Derived text can leak the underlying information.

    Proposed response

    Quarantine the page and align its audience with the most restrictive source.

Effort and cost — unknown

Setup work
  1. Define source boundaries, retention and decision-state vocabulary.

Ongoing review
  1. Resolve contradictions and remove withdrawn source material.

Unknowns
  1. Unknown: embedding, storage and refresh costs, permission fidelity and review load.

First check — proposed, unexecuted

Status

proposed_unexecuted

Synthetic fixture

Four invented work excerpts include an abandoned proposal, its replacement and two similarly named projects.

Acceptance checks
  1. Keep the projects separate.

  2. Cite the replacement decision and retain the abandoned status.

Stop conditions
  1. Real workplace content, live ingestion or access changes are required.

Limits

  1. Recommended privacy and review controls require separate validation; an indexed summary is not a new authority for workplace decisions.

Tasks: Email & messagesPlanning & operations

Useful for: Project managerResearcherSoftware developer

Editorial relevance from the title and source-supported actions; not evidence of the operator’s occupation.
Hermes AgentIndependent testing: No

Editorial candidate. This is source-backed reporting or a documented workflow. BotCurrent has not independently tested this task. A proposed Canadian adaptation is not evidence of a Canadian deployment or compliance certification.

Original source

https://engineering.razorpay.com/running-hermes-at-razorpay-a-network-isolated-self-improving-second-brain-for-every-employee-f91d56bea3f1 ↗

Attributed to: Ashwath Kumar, Razorpay Engineering

User report
Original action type
cross_app_investigation
What the source supports
Reports an employee’s scheduled ingestion of Slack, GitHub and Google Workspace into a personal wiki, entity graph and retrieval index.
Source outcome limits
Actions and outcomes are attributed user reports, not an independent reproduction or an execution guarantee.
Evidence attribution caveat
Actions and outcomes are attributed user reports, not an independent reproduction or an execution guarantee.
Source approval description
Deployment supports optional human gates for skill writes and risky commands; the example’s per-action configuration is unspecified.
Named tools
Hermes Agent, Kubernetes, Slack, GitHub, Google Workspace, Titan-v2 embeddings
Model or version
Unknown
Publication date
2026-07-12
Source last checked
2026-10-14
Demonstrated in Canada
Unknown
Canadian availability
Unknown
Proposed Canadian adaptation
Proposed only: scope employee consent, source access, retention, processor location and organizational privacy review before ingesting workplace records.
Recommended approval boundaries
Approve source scope and retention; review access changes and sharing beyond the employee’s private workspace.

More examples of this task

Source-backed approaches using this or another agent. Not independently tested.

Hermes Agent

Maintain a source-linked research wiki

Ingest approved sources, reconcile existing entity and concept pages, preserve provenance, update navigation and audit contradictions, stale claims and broken links.

Official recipe · Publication date unknown · Independently tested: No

https://github.com/NousResearch/hermes-agent/blob/main/skills/research/llm-wiki/SKILL.md ↗

This official recipe describes a procedure; it does not prove successful execution or independently reproduced results.

Source facts, caveats and approval gates
Source attribution
Hermes Agent; based on Karpathy’s LLM Wiki pattern
Source publication date
Unknown
Source checked
2026-10-14
Model or version
Unknown
Named tools
Hermes web/file tools; Markdown; content hashes; optional Obsidian
Source approval description
The human selects sources and domain; raw inputs are designated immutable. Discussion may be skipped in configured scheduled runs.
Evidence attribution caveat
This official recipe describes a procedure; it does not prove successful execution or independently reproduced results.
Source outcome limits
This official recipe describes a procedure; it does not prove successful execution or independently reproduced results.
Recommended approval gates
Approve corpus, storage and retention; review consequential interpretations and obtain permission before any new remote sync or source redistribution.
Canadian availability
Unknown
Demonstrated in Canada
Unknown
Proposed Canadian adaptation
Proposed: limit to authorized Canadian work material, assess cross-border processing and separate source excerpts from redistributable summaries.

OpenClaw

Build a searchable knowledge collection from videos and papers

The author taught OpenClaw to run installed video and paper tools, save condensed Markdown and index it for later retrieval.

User report · Published 2026-02-05 · Independently tested: No

https://notesfromzero.substack.com/p/building-knowledge-bases-for-my-openclaw ↗

Actions and outcomes are attributed user reports, not an independent reproduction or an execution guarantee.

Source facts, caveats and approval gates
Source attribution
Weijian Zhang
Source publication date
2026-02-05
Source checked
2026-10-14
Model or version
Unknown
Named tools
OpenClaw; Mac mini; youtube-summariser; paper-fluff-cutter; QMD.
Source approval description
The author installed packages and managed keys.
Evidence attribution caveat
Actions and outcomes are attributed user reports, not an independent reproduction or an execution guarantee.
Source outcome limits
Actions and outcomes are attributed user reports, not an independent reproduction or an execution guarantee.
Recommended approval gates
Approve source access, storage rights and any confidential documents before ingestion.
Canadian availability
Unknown
Demonstrated in Canada
Unknown
Proposed Canadian adaptation
Use permitted transcripts and papers, preserve source links and evaluate EN/FR retrieval. Avoid distributing copyrighted full texts.

Original plans extending this task

Proposed · Not tested. These plans do not add source-backed evidence.

A podcast back catalogue is too hard to search by topic