Task statement 5.6 · Context management and reliability

Preserve information provenance and handle uncertainty in multi-source synthesis

Three credible sources give three prices for battery packs: an industry survey says $139/kWh, a national lab $115 and a trade association $124. $139 is from December 2023, so it is simply older. $115 and $124 are both 2024 figures that genuinely disagree. The draft report says only "$124/kWh", with no source or date, so a reader cannot tell the two cases apart.

Scenario switchLike 5.3, this section runs on the exam guide's Multi-Agent Research System. The battery storage report from 5.3 is still being written. This time nothing failed: three sources were found, and they disagree.
Why it's tested

The tempting move is to give the reader one clean number. Picking the most recent or most authoritative-sounding value looks decisive, but it hides a disagreement the reader needed to weigh. Averaging $115 and $124 looks fair, but no source reported $119.50, so nothing can be cited for it. Passing findings on as tidy prose looks efficient, yet each summary can drop a claim's source and date, and no later agent can recover them.

KnowledgeWhere attribution is lost

Provenance is where a claim came from: its source, the passage and the date. In a multi-agent system it is lost at a summarisation step. When a subagent compresses its findings into prose, "pack prices are around $115–139 and falling" keeps the gist but drops which source said what, and when. Synthesis can only cite what reaches it, so the fix sits upstream: each subagent returns structured claim-source mappings, and every later agent preserves and merges them.

  1. Each subagent returns claim-source mappings, each with its publication or collection date.
  2. Document analysis completes with every value included, and annotates each conflict with the sources and methods instead of choosing.
  3. The coordinator decides how to reconcile: send a search for a third source, or pass both values on as contested. Here no comparable third source exists, so it passes both.
  4. Synthesis merges the mappings. It may shorten the wording, but every claim keeps its source and date.
  5. The report separates well-established from contested findings and gives each content type its natural form.

Interactive · three findings through synthesis

Top to bottom: three findings from different sources with prices and dates, what the synthesis subagent receives, the resulting report text, and two meters for claims traceable to a source and dates preserved FINDINGS FROM THE SEARCH AND DOCUMENT SUBAGENTS $139 / kWh industry survey published Dec 2023 $115 / kWh national lab report data collected 2024 $124 / kWh trade association data collected 2024 document analysis, then the coordinator WHAT THE SYNTHESIS SUBAGENT RECEIVES “Battery pack prices are around $115–139/kWh and falling.” 3 claims, each with source, excerpt and date — no sources, no dates, no methods — conflict flagged: $115 vs $124, both 2024, methods differ written into the report FINAL REPORT Battery pack prices fell to $124/kWh. Well established: prices fell. Both 2024 figures are below the survey's $139 (Dec 2023). Contested: the 2024 level. $115 (national lab, estimate from a bottom-up model) vs $124 (trade association, member survey). No third source found; neither preferred. one value chosen · no source · no date · no mention of disagreement CLAIMS TRACEABLE TO A SOURCE · CLAIMS 0 of 3 3 of 3 DATES PRESERVED · CLAIMS 0 of 3 3 of 3 The reader cannot tell the older figure from the genuine disagreement, or check either. The 2023 figure is older, not contradictory. Only the two 2024 figures genuinely conflict. Illustrative figures and sources.

Choose how findings travel between agents, and see what the report can still say about them.

Each prose summary kept the gist and dropped the sources. By the report, one number stands alone and nothing can be traced.

The mappings survive every handoff. The report can separate what is settled from what is contested, and name who said each.

Time reads downward: findings, the handoff to synthesis, the report. The sources do not change between the two runs. Only what travels with each claim does.

KnowledgeWhat one mapping carries

The mapping holds the claim, the source URL or document name, and the relevant excerpt, so a reader can check the claim. Two more fields stop two misreadings. The publication or collection date shows that $139 (2023) and $115 (2024) describe different years, so a temporal difference is not reported as a contradiction. The method keeps the source's methodological context, so a bottom-up estimate is not merged with a member survey. Here is the national lab's finding, with the conflict annotation from step 2.

// one finding, as the document analysis subagent returns it
{
  "claim": "Average battery pack price was $115/kWh",
  "source": "National lab annual storage cost report",
  "source_url": "https://example.org/storage-costs-2024.pdf",
  "excerpt": "…we estimate pack-level cost averaged $115/kWh…",
  "collected": "2024", "published": "2025-02",
  "method": "bottom-up cost model (estimate)",
  "conflicts_with": ["trade-association-2024"]
}

SkillAnnotate the conflict, then let the coordinator decide

Why did the draft say "$124/kWh" alone? Document analysis had been told to "return the best figure for each metric", so it chose the trade association's value during extraction, and the $115 figure never reached the coordinator. That is arbitrary selection, though it looked like judgement. The fix lets document analysis complete with both values included and the conflict annotated. The coordinator then decides, because only it can send another search. It asked web search for a third 2024 figure on a comparable method, found none, and passed both on as contested:

// coordinator → synthesis
{ "conflict": "2024 pack price",
  "values": ["national-lab-2024", "trade-association-2024"],
  "decision": "contested: no third source with a comparable method",
  "instruction": "report both with source and method; prefer neither" }

SkillA report that shows what is settled, in the right form

A finding is well established when the sources that address it agree, and contested when credible sources disagree for the same period and scope. A separate section for each lets the reader see what to rely on. The report also preserves each source's original characterisation: the lab calls $115 an estimate, so the report must not say "the lab found". Synthesis then renders each content type in its natural form rather than one uniform format. Financial data goes in a table, so values line up. News goes in prose, because an event needs its sequence and cause: as bullets, "cuts announced · lithium cheaper" no longer says which caused which. Technical findings go in a structured list, one checkable item each.

## Battery pack prices (illustrative figures and sources)
### Well established
Prices fell: both 2024 figures are below the Dec 2023 survey.
| Source            | Value    | Period   | Method                 |
|-------------------|----------|----------|------------------------|
| Industry survey   | $139/kWh | Dec 2023 | buyer survey           |
| National lab      | $115/kWh | 2024     | bottom-up model (est.) |
| Trade association | $124/kWh | 2024     | member survey          |
### Contested
The 2024 level: the lab estimates $115; the association reports $124.
The methods differ and no third source was found. Neither is preferred.
### Recent news
In February 2025 two cell makers announced further price cuts, citing
cheaper lithium (trade press, 2025-02).
### Technical findings
- Chemistry: LFP in 9 of 12 projects reviewed (national lab, 2024)
- Pack share of installed system cost: 41% (national lab, 2024)

What this looks like in the platform. With citations: {"enabled": true} on a document block, the Claude API returns each supporting passage in cited_text, which the docs say is "guaranteed to contain valid pointers to the provided documents" (checked 2026-10-01). The guide does not name this feature; it tests that the mapping survives every handoff.

For each figure in the report, ask: can it be traced to a source and date, and do credible sources for the same period agree?
Traced; sources agree or differ only by dateWell established. Show dated values as a trend.
Traced; sources disagree for the same periodContested. Show each value with source and method.
Not tracedA mapping was dropped upstream. Restore it there.
Takeaway
  • Attribution is lost where findings are compressed into prose, so claim-source mappings with dates must survive every handoff.
  • Document analysis includes and annotates conflicting values; the coordinator decides whether to seek another source or pass them on as contested.
  • Financial data as tables, news as prose, technical findings as structured lists.

check yourself

0 / 3 answered

The research system is writing an offshore wind report for a client's board. Five sources agree that auction prices fell from 2021 to 2024. Two government agencies give 2024 capacity factors of 38% and 45%, measured on different turbine fleets. The board wants to see which conclusions it can rely on without opening the sources.Which report structure best meets the board's request?

Select one.

Document analysis is processing 30 reports for a study of hospital nurse staffing. Two credible 2024 reports give nurse-to-patient ratios of 1:5 and 1:7 for the same region. Team policy says a conflict gets a third source when one exists, and only the coordinator can dispatch searches.Which way of handling the conflict best follows the policy while keeping the analysis moving?

Select one.

A report on chip supply combines quarterly revenue for five firms, three news events from the past month, and benchmark results for two chip designs. Synthesis writes every section as bullets. Readers say the revenue figures are hard to compare and the news items no longer explain why events happened. Every claim must keep its own citation.Which TWO changes best fix what readers reported while keeping every citation?

Select two.

sources

  1. Anthropic Engineering, "How we built our multi-agent research system": subagents condense findings for the lead agent, and a CitationAgent "processes the documents and research report to identify specific locations for citations". Checked 2026-10-01.
  2. Claude Platform Docs, "Citations": citations: {"enabled": true} on document blocks, cited_text, and "guaranteed to contain valid pointers to the provided documents". Checked 2026-10-01.
  3. Claim-source mappings, conflict annotation with coordinator reconciliation, temporal data, well-established versus contested sections, and content-type rendering: Claude Certified Architect — Foundations Exam Guide v1.0, Domain 5.6. Scenario: the guide's Multi-Agent Research System. Figures, sources and news items are illustrative.