Agents access ASD through 19+ MCP tools organized by category.
| Tool | Description |
|---|
code_query | Search the semantic index by symbol name or pattern |
code_read | Read a symbol’s source with surrounding context |
code_search | Full-text search across the indexed codebase |
callers_of | Return all callers of a symbol (inbound edges) |
callees_of | Return all callees of a symbol (outbound edges) |
| Tool | Description |
|---|
ledger_get | Get all entries for a symbol |
ledger_find | Search entries by tag, text, or kind |
ledger_append | Append a new entry (decision, hazard, rationale, constraint) |
ledger_approve | Approve a pending entry |
ledger_reject | Reject a pending entry |
ledger_withdraw | Withdraw an entry |
ledger_supersede | Replace an entry with a new one |
ledger_rebind | Re-attach an entry to a new qname (after a rename) |
| Tool | Description |
|---|
effect_declare | Declare an effect on a symbol |
effects_of | Get declared effects for a symbol |
traces_of | Get the transitive effect trace for a symbol |
| Tool | Description |
|---|
think_speculate | Record a hypothesis with a confidence score |
think_model | Record a mental model spanning multiple symbols |
think_failed | Record a failed attempt (what was tried, why it failed) |
think_question | Record an open question blocking confident action |
think_list | List thinking entries, filtered by kind or symbol |
| Tool | Description |
|---|
health | Check index freshness and server status |
reindex | Trigger a full reindex of the workspace |
audit_tail | Stream recent audit events |
audit_verify | Verify the hash chain of the audit log |