dsh-action-ledger
Run the following command in DeepSeek Harness:
dsh plugin install MkaliezZ/dsh-action-ledger
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install MkaliezZ/dsh-action-ledger in DeepSeek Harness to install this plugin; the source code is available at https://github.com/MkaliezZ/dsh-action-ledger.
About this plugin
DSH sessions accumulate a stream of discrete machine facts around every tool call—the request, the approval state, the result—yet none of that forms a single human-readable timeline. AgentFuse answers whether a call may proceed; dsh-action-ledger answers which lifecycle states DSH durably recorded around that call. It is a read-only projection layered over existing DSH session facts, introducing no new standalone event types and requiring no patching of the generated event catalog.
The projected ledger unfolds in deterministic order: requested, approval-requested, approved or denied, completed or failed. Output is strictly bounded, raw tool arguments are never copied into the ledger, no tool dispatch is triggered, and no source mutation occurs. When a lifecycle node is absent from DSH facts, the ledger does not fabricate it; it only presents the nodes DSH actually recorded and makes no inferences beyond those facts.
It is built for engineers who build or debug agent workflows on DSH and need to audit the full approval-and-execution trail of a tool call, or replay what actually happened rather than what was merely permitted. It is not a replacement for DSH session persistence and not an authorization system; it is a faithful, bounded mirror of the facts already on record.
Use Cases
- Replay the full approval-and-execution trail of a tool call while debugging agent workflows
- Audit the durable action facts in a DSH session to confirm which lifecycle nodes are present
- Embed a human-readable call timeline in a report without parsing raw event streams
Best For
- Platform engineers building and debugging agent workflows on DSH
- Security auditors who need the full lifecycle trail of a tool call beyond a permission verdict
- DevOps or SRE engineers rendering human-readable summaries from DSH session event streams
Related Plugins
A method pack that makes AI coding agents plan against your real baseline, prove completion with fresh evidence, and reduce reworks and unsafe changes.
Turns the DeepSeek Harness session into a captain that builds a durable sub-agent team, splits goals into dependency-aware tasks, and coordinates work via direct messages and a live Web UI.
Gives coding agents design judgment, letting Claude Code, Cursor, and 70+ agents generate and iterate high-quality UI, presentations, and graphics right from the terminal.
Run the Pi ecosystem's plugins on DeepSeek Harness, unmodified, via a compatibility layer that implements Pi's public extension ABI on DSH's native services.