Deep Dive

# Federal Contract Opportunity Deep Dive

## User Input
- **Target opportunity:** [Solicitation number, GovTribe link, or opportunity title plus agency]

## Goal
Use GovTribe MCP tools to resolve a specific federal contract opportunity, notice, solicitation, sources sought, presolicitation, or related contracting notice and produce a grounded record-level brief.
Summarize the key notice facts, surface the most useful government files, trace the opportunity to directly linked awards or incumbent signals when the evidence supports that connection, and include parent vehicle or IDV context when the resolved notice clearly references it.

## Required Input
The user must provide a specific target opportunity before analysis begins.

Accept any of the following:
- Solicitation number
- GovTribe link
- Opportunity title plus agency
- Plain-language description only if it is specific enough to resolve a single opportunity

Input rules:
- If the input resolves cleanly to one target, proceed immediately.
- If the input is too vague to resolve a single opportunity, ask for the minimum missing detail needed to proceed.
- Government files, direct linked-award or incumbent trace, and vehicle or IDV context are part of the default workflow when the retrieved evidence supports them.
- Do not guess the target.
- Do not start substantive analysis until the target is resolved.

## Workflow

### Steps
1. Call `Documentation` once with `collections=["govtribe-for-agents"]`, `article_names=["Choose a search mode and write queries", "Manage search context", "Filter by related records and hierarchies", "Aggregations and leaderboards", "Vector-store content retrieval", "Troubleshoot search results"]`, and `max_tokens=12000`.
    - Treat the returned guide articles as binding for search behavior, query construction, relationship filters, direct-link aggregation review, vector-store retrieval, search-context management, and troubleshooting used in this workflow.
    - Use the selected tool schema as the source of truth for tool-specific arguments, available fields, relationship fields, filters, sorts, aggregation keys, and response shapes.

2. Resolve the target opportunity with `Search_Federal_Contract_Opportunities`.
    - Favor exact lookup first.
    - For exact solicitation numbers, quoted titles, or exact GovTribe-derived identifiers, use a double quoted `"query"`.
    - If the user provides a GovTribe link, use the record identity embedded in the link when possible; otherwise resolve it through exact quoted lookup.
    - If the user provides an opportunity title plus agency, use the strongest exact title phrase plus agency context, then disambiguate with returned fields.
    - Use `fields_to_return` explicitly. At minimum request:
        - `govtribe_id`, `govtribe_url`, `govtribe_type`, `solicitation_number`, `name`, `opportunity_type`, `opportunity_state`, `part_of_mas`, `set_aside_type`, `posted_date`, `due_date`, `award_date`, `descriptions`, `govtribe_ai_summary`, `federal_meta_opportunity_id`, `federal_contract_vehicle`, `federal_agency`, `place_of_performance`, `naics_category`, `psc_category`, `government_files`, `federal_contract_awards`, `federal_contract_idvs`, `points_of_contact`
    - Use date sorting only when multiple historical variants or amendment-related versions need review.

3. If multiple opportunities match, disambiguate with the minimum additional evidence needed.
    - Compare agency, opportunity type, posted date, due date, set-aside, NAICS, PSC, and vehicle context.
    - Do not merge multiple possible matches into one narrative.
    - If the record still cannot be resolved to a single opportunity, stop and ask for one clarifying detail.

4. If exact resolution fails and the user input is still specific enough, run one narrow recovery retry before asking for clarification.
    - Broaden only one dimension at a time, such as title phrasing or agency context.
    - If the retry still does not produce a single defensible match, stop and ask for the minimum missing detail.

5. Pull government attachments tied to the resolved opportunity with `Search_Government_Files`.
    - Use `federal_contract_opportunity_ids` with the resolved opportunity GovTribe ID.
    - Request:
        - `govtribe_id`, `govtribe_url`, `name`, `posted_date`, `content_snippet`, `download_url`, `parent_record`, `govtribe_ai_summary`
    - Focus on files that materially explain the scope, amendments, Q&A, statements of work, attachments, or evaluation context.
    - If no files are returned, say that clearly instead of implying the notice had no attachments.

6. Use only direct linkage evidence for linked-award, incumbent, and post-award context.
    - If the resolved opportunity has `federal_meta_opportunity_id`, call `Search_Federal_Contract_Awards` with `federal_meta_opportunity_ids`.
    - Request:
        - `govtribe_id`, `govtribe_url`, `name`, `contract_number`, `award_date`, `completion_date`, `ultimate_completion_date`, `contract_type`, `descriptions`, `govtribe_ai_summary`, `dollars_obligated`, `ceiling_value`, `set_aside_type`, `awardee`, `parent_of_awardee`, `federal_contract_idv`, `federal_contract_vehicle`, `contracting_federal_agency`, `funding_federal_agency`, `naics_category`, `psc_category`, `place_of_performance`, `originating_federal_meta_opportunity_id`, `originating_federal_contract_opportunity`
    - If multiple linked awards are returned and concentration or post-award interpretation matters, add an aggregation-only pass on the same direct linkage:
        - Use `Search_Federal_Contract_Awards`
        - Use `per_page: 0`
        - Reuse the same `federal_meta_opportunity_ids`
        - Use `aggregations` such as `dollars_obligated_stats`, `top_awardees_by_dollars_obligated`, and `top_contracting_federal_agencies_by_dollars_obligated`
    - Use that aggregation branch only to summarize the direct historical award footprint tied to the notice thread and to distinguish a single obvious incumbent from a more distributed linked-award history.
    - Only describe a linked award, incumbent, or post-award outcome when the connection is direct and returned by the dataset.
    - Do not infer incumbent status from loose similarity, keywords, or adjacent awards.

7. Normalize awardee identity only when a linked award was found and identity context materially improves the answer.
    - If awardee identity, parent-child structure, or legal-entity naming matters, use `Search_Vendors`.
    - Prefer `vendor_ids` from award relationships when available.
    - If no vendor ID is available, use an exact quoted company name in `Search_Vendors.query`.
    - Request `govtribe_id`, `govtribe_url`, `name`, `uei`, `dba`, `parent_or_child`, `parent`, `business_types`, `sba_certifications`, and `govtribe_ai_summary`.
    - Use vendor normalization to clarify identity, not to replace opportunity or award evidence.

8. If the opportunity clearly references a vehicle or IDV, run one targeted parent lookup.
    - IDV path with `Search_Federal_Contract_IDVs`:
        - Use the linked IDV when available, or an exact quoted contract number if that is all you have.
        - Request `govtribe_id`, `govtribe_url`, `name`, `contract_number`, `award_date`, `last_date_to_order`, `contract_type`, `pricing_type`, `description`, `govtribe_ai_summary`, `ceiling_value`, `set_aside`, `solicitation_procedures`, `extent_competed`, `legislative_mandate`, `multiple_or_single_award`, `awardee`, `parent_of_awardee`, `federal_contract_vehicle`, `contracting_federal_agency`, `funding_federal_agency`, `naics_category`, `psc_category`, `place_of_performance`, `task_orders`, `blanket_purchase_agreements`, `originating_federal_meta_opportunity_id`, and `originating_federal_contract_opportunity`
    - Vehicle path with `Search_Federal_Contract_Vehicles`:
        - Use the linked vehicle when available, or an exact quoted vehicle name if needed.
        - Request `govtribe_id`, `govtribe_url`, `name`, `award_date`, `last_date_to_order`, `contract_type`, `descriptions`, `govtribe_ai_summary`, `set_aside_type`, `shared_ceiling`, `originating_federal_meta_opportunity_id`, `originating_federal_contract_opportunity`, `federal_agency`, and `federal_contract_awards`
    - Do not call both parent tools unless the evidence requires both and the extra call materially improves correctness.

9. Escalate file-content retrieval only when it is clearly necessary.
    - If `content_snippet` is not enough and the user explicitly needs deeper document grounding, use `Add_To_Vector_Store`, then `Search_Vector_Store`.
    - State clearly when no files, no linked awards, no parent context, or no confirmed incumbent can be shown from the retrieved record set.

## Output Format
Use markdown tables for key facts, key files, and any multi-row linked-award summaries.

Return the answer in this order:

1. **Target Opportunity Summary**
    - Briefly explain how the opportunity was resolved

2. **Key Notice Facts**
    - Use a required 2-column markdown table for the key facts
    - Include solicitation number, opportunity type, agency, posted and due dates, set-aside, NAICS/PSC, place of performance, vehicle context, and the most relevant scope language

3. **Key Government Files**
    - Use a compact markdown table
    - Recommended columns: `File`, `Posted`, `Why It Matters`

4. **Linked Awards / Incumbent Trace**
    - Include only direct linked-award evidence
    - If multiple linked awards were retrieved, you may use a compact markdown table
    - If the linked-award aggregation branch was used, include a short aggregated history note
    - If no direct link is confirmed, say that clearly

5. **Parent Vehicle or IDV Context**
    - Include only if applicable and actually retrieved

6. **Risks, Gaps, or Unknowns**
    - Briefly note identity ambiguity, missing attachments, unclear linkage, or other data limits

7. **Overall Confidence**
    - State overall confidence and why

## Citation Rules
- Only cite sources retrieved in the current workflow.
- Never fabricate citations, URLs, IDs, or quote spans.
- Use exactly the citation format required by the host application.
- Attach citations to the specific claims they support, not only at the end.

## Grounding Rules
- Base claims only on provided context or GovTribe MCP tool outputs.
- If sources conflict, state the conflict explicitly and attribute each side.
- If the context is insufficient or irrelevant, narrow the answer or state that the goal cannot be fully completed from the available evidence.
- If a statement is an inference rather than a directly supported fact, label it as an inference.

# Federal Contract Award Deep Dive

## User Input
- **Target award:** [Contract number, PIID, order number, GovTribe link, or award title plus agency]

## Goal
Use GovTribe MCP tools to resolve a specific federal contract award, task order, BPA call, or IDV-linked order and produce a grounded record-level brief.
Summarize the key contract facts, trace the award back to the originating notice when possible, surface the most useful government files, and explain modification or lifecycle evidence.

## Required Input
The user must provide a specific target award before analysis begins.

Accept any of the following:
- Contract number or PIID
- Order number
- GovTribe link
- Award title plus agency
- Plain-language description only if it is specific enough to resolve a single award

Input rules:
- If the input resolves cleanly to one target, proceed immediately.
- If the input is too vague to resolve a single award, ask for the minimum missing detail needed to proceed.
- Do not guess the target.
- Do not start substantive analysis until the target is resolved.

## Workflow
### Rules

### Steps
1. Call `Documentation` once with `collections=["govtribe-for-agents"]`, `article_names=["Choose a search mode and write queries", "Manage search context", "Filter by related records and hierarchies", "Vector-store content retrieval", "Troubleshoot search results"]`, and `max_tokens=12000`.
    - Treat the returned guide articles as binding for search behavior, query construction, relationship filters, vector-store retrieval, search-context management, and troubleshooting used in this workflow.
    - Use the selected tool schema as the source of truth for tool-specific arguments, available fields, relationship fields, filters, sorts, and response shapes.

2. Resolve the target award with `Search_Federal_Contract_Awards`.
    - Favor exact lookup first.
    - For exact contract numbers, PIIDs, order numbers, quoted titles, or exact GovTribe-derived identifiers, use a double quoted `"query"`.
    - If the user provides a GovTribe link, use the record identity embedded in the link when possible; otherwise resolve it through exact quoted lookup.
    - If the user provides an award title plus agency, use the strongest exact title phrase plus agency context, then disambiguate with returned fields.
    - Use `fields_to_return` explicitly. At minimum request:
        - `govtribe_id`, `govtribe_url`, `govtribe_type`, `name`, `contract_number`, `award_date`, `completion_date`, `ultimate_completion_date`, `contract_type`, `pricing_type`, `set_aside_type`, `solicitation_procedures`, `research_type`, `legislative_mandate`, `extent_competed`, `number_of_offers_received`, `descriptions`, `govtribe_ai_summary`, `dollars_obligated`, `ceiling_value`, `awardee`, `parent_of_awardee`, `federal_contract_idv`, `federal_contract_vehicle`, `contracting_federal_agency`, `funding_federal_agency`, `naics_category`, `psc_category`, `place_of_performance`, `transaction_contacts`, `originating_federal_meta_opportunity_id`, `originating_federal_contract_opportunity`
    - Use date sorting only when multiple historical variants or similarly numbered records need review.

3. If multiple awards match, disambiguate with the minimum additional evidence needed.
    - Compare agency, awardee, award date, contract type, obligation size, and vehicle context.
    - Do not merge multiple possible matches into one narrative.
    - If the record still cannot be resolved to a single award, stop and ask for one clarifying detail.

4. If exact resolution fails and the user input is still specific enough, run one narrow recovery retry before asking for clarification.
    - Broaden only one dimension at a time, such as title phrasing or agency context.
    - If the retry still does not produce a single defensible match, stop and ask for the minimum missing detail.

5. Normalize awardee identity only when it materially improves the answer.
    - If awardee identity, parent-child structure, or legal-entity naming matters, use `Search_Vendors`.
    - Prefer `vendor_ids` from award relationships when available.
    - If no vendor ID is available, use an exact quoted company name in `Search_Vendors.query`.
    - Request `govtribe_id`, `govtribe_url`, `name`, `uei`, `dba`, `parent_or_child`, `parent`, `business_types`, `sba_certifications`, and `govtribe_ai_summary`.
    - Use vendor normalization to clarify identity, not to replace award evidence.

6. Trace the originating notice or opportunity as a required branch.
    - If `originating_federal_contract_opportunity.govtribe_id` is already available from the award result, use it directly as the starting point for this branch rather than treating the trace as already complete.
    - If `originating_federal_meta_opportunity_id` is available from the award result or linked evidence, resolve it with `Search_Federal_Contract_Opportunities` using `federal_meta_opportunity_ids`.
    - If the direct relationship is absent but the award still provides an exact identifier or strong context, run an independent opportunity search before concluding that no originating opportunity can be confirmed.
    - Use this recovery order:
        1. ID-defined lookup with `federal_contract_opportunity_ids` or `federal_meta_opportunity_ids` when those IDs are available.
        2. Exact quoted lookup using the best available solicitation number, order number, contract number, or exact title phrase.
        3. Retries that broaden only one dimension each, such as agency context or title phrasing, if the first independent lookup does not confirm a single defensible opportunity.
    - If the award clearly sits under a linked IDV or vehicle and direct opportunity tracing still fails, use one targeted parent lookup early if that parent record may expose `originating_federal_contract_opportunity`.
    - Use exact or tightly scoped opportunity resolution before broader retries.
    - Request opportunity fields only when needed:
        - `govtribe_id`, `govtribe_url`, `solicitation_number`, `name`, `opportunity_type`, `posted_date`, `due_date`, `descriptions`, `govtribe_ai_summary`, `federal_meta_opportunity_id`, `federal_agency`, `government_files`, `points_of_contact`
    - If more than one plausible opportunity is returned, disambiguate with agency, title, solicitation number, dates, file trail, and relationship context before concluding.
    - Do not stop at intent, intermediate reasoning, or a partial retry. Finish this branch with either one confirmed originating opportunity or an explicit statement that the available searches did not confirm one.
    - If no originating opportunity can be confirmed, say that clearly and do not infer a notice thread from loose keyword similarity.

7. Pull government attachments with `Search_Government_Files` only after an originating opportunity has been confirmed.
    - Use `federal_contract_opportunity_ids` with the resolved opportunity GovTribe ID.
    - Request `govtribe_id`, `govtribe_url`, `name`, `posted_date`, `content_snippet`, `download_url`, `parent_record`, and `govtribe_ai_summary`.
    - Focus on files that materially explain the scope, amendments, Q&A, statements of work, attachments, or evaluation context.
    - If no originating opportunity is confirmed, state that there is no confirmed opportunity-file trail.

8. Pull modification and lifecycle evidence with `Search_Federal_Transactions`.
    - Use the resolved `federal_contract_award_ids`.
    - Request `govtribe_id`, `date`, `last_mod_number`, `reason_for_modification`, `total_value`, `federal_value`, `awardee`, `funding_federal_agency`, and `contracting_federal_agency`.
    - Use `sort` with `transactionDate` descending.
    - Treat transaction rows as supporting lifecycle evidence such as option exercises, funding increments, administrative changes, or other modifications.
    - Do not infer option structure, extension status, or recompete timing unless the returned award or transaction evidence actually supports it.

9. If the award clearly sits under an IDV or vehicle, run one targeted parent lookup.
    - IDV path with `Search_Federal_Contract_IDVs`:
        - Use the linked IDV when available, or an exact quoted contract number if that is all you have.
        - Request `govtribe_id`, `govtribe_url`, `name`, `contract_number`, `award_date`, `last_date_to_order`, `contract_type`, `pricing_type`, `description`, `govtribe_ai_summary`, `ceiling_value`, `set_aside`, `solicitation_procedures`, `extent_competed`, `legislative_mandate`, `multiple_or_single_award`, `awardee`, `parent_of_awardee`, `federal_contract_vehicle`, `contracting_federal_agency`, `funding_federal_agency`, `naics_category`, `psc_category`, `place_of_performance`, `task_orders`, `blanket_purchase_agreements`, `originating_federal_meta_opportunity_id`, and `originating_federal_contract_opportunity`
    - Vehicle path with `Search_Federal_Contract_Vehicles`:
        - Use the linked vehicle when available, or an exact quoted vehicle name if needed.
        - Request `govtribe_id`, `govtribe_url`, `name`, `award_date`, `last_date_to_order`, `contract_type`, `descriptions`, `govtribe_ai_summary`, `set_aside_type`, `shared_ceiling`, `originating_federal_meta_opportunity_id`, `originating_federal_contract_opportunity`, `federal_agency`, and `federal_contract_awards`
    - Do not call both parent tools unless the evidence requires both and the extra call materially improves correctness.

10. Escalate file-content retrieval only when it is clearly necessary.
    - If `content_snippet` is not enough and the user explicitly needs deeper document grounding, use `Add_To_Vector_Store`, then `Search_Vector_Store`.
    - State clearly when the originating notice, file trail, parent lineage, or lifecycle evidence cannot be confirmed from the retrieved record set.

## Output Format
Use markdown tables for key facts, key files, and any multi-row modification history.

Return the answer in this order:

1. **Target Award Summary**
    - Briefly explain how the award was resolved

2. **Key Contract Facts**
    - Use a required 2-column markdown table for the key facts
    - Include contract number, awardee, agencies, contract type, dates, dollars, set-aside, NAICS/PSC, place of performance, and any directly linked vehicle or IDV fields that were retrieved

3. **Originating Notice / Opportunity Trace**
    - State the linked opportunity if confirmed
    - If not confirmed, say that clearly

4. **Key Government Files**
    - Use a compact markdown table
    - Recommended columns: `File`, `Posted`, `Why It Matters`
    - If no confirmed opportunity-file trail exists, say that clearly instead of presenting an empty table

5. **Modification and Lifecycle Notes**
    - If more than one transaction or lifecycle event is surfaced, use a compact markdown table
    - Recommended columns: `Date`, `Modification`, `Why It Matters`
    - If only one event matters, a short prose note is fine

6. **Parent Vehicle or IDV Context**
    - Include only if applicable and actually retrieved

7. **Risks, Gaps, or Unknowns**
    - Briefly note identity ambiguity, missing file trail, sparse mod history, or other data limits

8. **Overall Confidence**
    - State overall confidence and why

## Citation Rules
- Only cite sources retrieved in the current workflow.
- Never fabricate citations, URLs, IDs, or quote spans.
- Use exactly the citation format required by the host application.
- Attach citations to the specific claims they support, not only at the end.

## Grounding Rules
- Base claims only on provided context or GovTribe MCP tool outputs.
- If sources conflict, state the conflict explicitly and attribute each side.
- If the context is insufficient or irrelevant, narrow the answer or state that the goal cannot be fully completed from the available evidence.
- If a statement is an inference rather than a directly supported fact, label it as an inference.

# Vendor Deep Dive

## User Input
- **Target vendor:** [Vendor name, UEI, CAGE, GovTribe link, or description]

## Goal
Use GovTribe MCP tools to resolve a target vendor and produce a grounded government activity briefing across the vendor's recent federal contract awards, federal contract IDVs, federal grant awards, state and local contract awards, and government-related news signals.

## Required Input
The user must provide a **target vendor** before analysis begins.

Accept any of the following:
- Company name
- UEI
- CAGE
- GovTribe link
- Plain-language description only if it resolves to a single vendor without ambiguity

Input rules:
- The workflow always analyzes the resolved vendor's last 5 years of federal contract awards, federal contract IDVs, federal grant awards, and state and local contract awards.
- The workflow always analyzes a bounded recent window of government-related news, typically the last 24 months.
- Derive agencies, customers, capabilities, classifications, assistance types, vehicles, IDVs, and value bands from the returned evidence rather than treating them as user-selectable modes.
- Treat any additional context as interpretation context only. Do not let it turn this workflow into family roll-up analysis, subaward analysis, competitor ranking, buyer expansion planning, or past-performance matching.
- If the input resolves cleanly to one vendor, proceed immediately.
- If the input is too vague or ambiguous, ask for the minimum missing detail needed to proceed.
- Do not guess the target vendor.
- Do not start substantive analysis until the vendor scope is resolved.

## Workflow

### Steps
1. Call `Documentation` once with `collections=["govtribe-for-agents"]`, `article_names=["Choose a search mode and write queries", "Manage search context", "Date filtering", "Filter by related records and hierarchies", "Aggregations and leaderboards", "Troubleshoot search results"]`, and `max_tokens=12000` before any other GovTribe tool.
   - Treat the returned guide articles as binding for search behavior, query construction, date filtering, relationship filters, aggregation-first review, search-context management, and troubleshooting used in this workflow.
   - Use the selected tool schema as the source of truth for tool-specific arguments, available fields, filters, sorts, aggregation keys, and response shapes.
2. Resolve the target vendor identity explicitly with `Search_Vendors`.
   - Use `Search_Vendors` for company name, GovTribe link, UEI, CAGE, or other known vendor identity.
   - For exact names or identifiers, use a double quoted `"query"`.
   - Request `fields_to_return` explicitly. At minimum request `govtribe_id`, `govtribe_url`, `name`, `uei`, `dba`, `cage_codes`, `activation_date`, `registration_date`, `registration_deleted_date`, `registration_expiration_date`, `business_types`, `sba_certifications`, `parent_or_child`, `parent`, `naics_category`, and `govtribe_ai_summary`.
   - When registration status or certification timing matters, also request the relevant vendor certification lifecycle fields exposed on the row, including `sba_cert_8a_expiration_date`, `sba_cert_8a_joint_venture_expiration_date`, `sba_cert_hubzone_expiration_date`, `sba_cert_small_disadvantaged_business_expiration_date`, `sba_cert_sdvosb_expiration_date`, and `sba_cert_vosb_expiration_date`.
   - Normalize the legal name, common name or DBA, UEI when available, CAGE only when it appears in retrieved evidence, and any parent or subsidiary relationships returned by the tool.
   - Do not invent a direct `cage` filter because `Search_Vendors` does not document one.
3. Lock the exact vendor scope before cross-dataset search.
   - State the resolved legal name and any confirmed DBA names or naming variants that are safe to reuse in later search steps.
   - Keep parent or child entities out of scope unless the returned evidence shows the user clearly meant one of them instead of the resolved vendor.
   - If the vendor resolution is still ambiguous after one careful retry, ask for clarification and stop.
4. Run the federal contract-award path over the last 5 years.
   - Use `Search_Federal_Contract_Awards`.
   - Start with one aggregation-first pass using `vendor_ids` and an `award_date_range` covering the last 5 years.
   - Use `per_page: 0` and `aggregations` such as `dollars_obligated_stats`, `top_funding_federal_agencies_by_dollars_obligated`, `top_contracting_federal_agencies_by_dollars_obligated`, `top_naics_codes_by_dollars_obligated`, `top_psc_codes_by_dollars_obligated`, `top_federal_contract_vehicles_by_dollars_obligated`, `top_idvs_by_dollars_obligated`, `top_set_aside_types_by_dollars_obligated`, and `top_locations_by_dollars_obligated`.
   - Then retrieve representative rows with the same `vendor_ids` and the same 5-year `award_date_range`.
   - Add only supported filters when returned evidence makes them necessary for cleanup or follow-on interpretation, such as `contracting_federal_agency_ids`, `funding_federal_agency_ids`, `naics_category_ids`, `psc_category_ids`, `federal_contract_vehicle_ids`, `federal_contract_idv_ids`, `federal_contract_award_types`, and `set_aside_types`.
   - Request `fields_to_return` explicitly. At minimum request `govtribe_id`, `govtribe_url`, `name`, `contract_number`, `award_date`, `completion_date`, `ultimate_completion_date`, `contract_type`, `descriptions`, `govtribe_ai_summary`, `dollars_obligated`, `ceiling_value`, `set_aside_type`, `awardee`, `parent_of_awardee`, `contracting_federal_agency`, `funding_federal_agency`, `naics_category`, `psc_category`, `federal_contract_vehicle`, `federal_contract_idv`, `place_of_performance`, `originating_federal_meta_opportunity_id`, and `originating_federal_contract_opportunity`.
   - Use contract-award text fields such as `descriptions` and `govtribe_ai_summary` to interpret scope, but do not let text outrank the structured vendor match.
   - Use `search_mode: "semantic"` only when the first pass is too noisy or too thin to explain the vendor's main federal contract lanes.
5. Run the federal contract IDV path over the last 5 years.
   - Use `Search_Federal_Contract_IDVs`.
   - Start with one aggregation-first pass using `vendor_ids` and an `award_date_range` covering the last 5 years.
   - Use `per_page: 0` and `aggregations` such as `top_awardees_by_doc_count`, `top_funding_federal_agencies_by_doc_count`, `top_contracting_federal_agencies_by_doc_count`, `top_vehicles_by_doc_count`, `top_naics_codes_by_doc_count`, `top_psc_codes_by_doc_count`, `top_set_aside_types_by_doc_count`, and `top_transaction_contacts_by_doc_count`.
   - Then retrieve representative rows with the same `vendor_ids` and the same 5-year `award_date_range`.
   - Request `fields_to_return` explicitly. At minimum request `govtribe_id`, `govtribe_url`, `name`, `contract_number`, `award_date`, `last_date_to_order`, `ceiling_value`, `contract_type`, `pricing_type`, `multiple_or_single_award`, `awardee`, `parent_of_awardee`, `contracting_federal_agency`, `funding_federal_agency`, `naics_category`, `psc_category`, `federal_contract_vehicle`, `originating_federal_meta_opportunity_id`, `originating_federal_contract_opportunity`, `place_of_performance`, `set_aside`, `solicitation_procedures`, `extent_competed`, `legislative_mandate`, and `transaction_contacts`.
   - Only if IDV rows show vehicle concentration that materially sharpens the analysis, follow the strongest returned `federal_contract_vehicle` relationship IDs into `Search_Federal_Contract_Vehicles`.
   - Use `search_mode: "semantic"` only when the first pass is too noisy or too thin to explain the vendor's main IDV footprint.
6. Run the federal grant-award path over the last 5 years.
   - Use `Search_Federal_Grant_Awards`.
   - Start with one aggregation-first pass using `vendor_ids` and an `award_date_range` covering the last 5 years.
   - Use `per_page: 0` and `aggregations` such as `dollars_obligated_stats`, `top_funding_federal_agencies_by_dollars_obligated`, `top_awardees_by_dollars_obligated`, `top_locations_by_dollars_obligated`, and `top_federal_grant_programs_by_dollars_obligated`.
   - Then retrieve representative rows with the same `vendor_ids` and the same 5-year `award_date_range`.
   - Request `fields_to_return` explicitly. At minimum request `govtribe_id`, `govtribe_url`, `name`, `description`, `grant_number`, `award_date`, `completion_date`, `ultimate_completion_date`, `assistance_type`, `assistance_recipient_type`, `dollars_obligated`, `awardee`, `parent_of_awardee`, `funding_federal_agency`, `contracting_federal_agency`, `federal_grant_program`, `place_of_performance`, and `govtribe_ai_summary`.
   - Use `search_mode: "semantic"` only when the first pass is too noisy or too thin to explain the vendor's main grant footprint.
7. Run the state and local contract-award path over the last 5 years.
   - Use `Search_State_And_Local_Contract_Awards`.
   - Because this dataset does not support `vendor_ids`, build the initial `query` from the resolved legal name and only the safest confirmed DBA variants.
   - Start with one aggregation-first pass using the strongest exact vendor-name query you can support and an `award_date_range` covering the last 5 years.
   - Use `per_page: 0` and `aggregations` such as `dollars_obligated_stats`, `top_contract_entities_by_dollars_obligated`, `top_states_by_dollars_obligated`, `top_nigp_codes_by_dollars_obligated`, and `top_unspsc_codes_by_dollars_obligated`.
   - Then retrieve representative rows with the same 5-year `award_date_range` and the same exact-name query.
   - Request `fields_to_return` explicitly. At minimum request `govtribe_id`, `govtribe_url`, `name`, `description`, `awardee_name`, `contracting_organization`, `award_date`, `completion_date`, `estimated_annual_value`, `contract_amount`, `cumulative_value`, `contract_number`, `contract_type`, `state`, `nigp_categories`, `unspsc_categories`, `government_files`, `points_of_contact`, and `govtribe_ai_summary`.
   - Exclude ambiguous name matches aggressively. If the tool returns noisy results and you cannot defend attribution to the resolved vendor, say the state and local surface is inconclusive instead of overclaiming.
   - Use `search_mode: "semantic"` only for one careful fallback pass when exact-name retrieval is too thin and you still have a defensible vendor-specific query.
8. Run the government-related news path over a recent window.
   - Use `Search_Government_Related_News_Articles`.
   - Start with a bounded recent window such as the last 24 months in `date_published`.
   - Build the initial `query` from the resolved legal name and only the safest confirmed DBA variants.
   - Request `fields_to_return` explicitly. At minimum request `govtribe_id`, `govtribe_url`, `title`, `subheader`, `published_date`, `site_name`, and `body`.
   - Sort by `datePublished` descending for chronology-first briefing, or by `_score` only when one careful semantic follow-on search materially improves relevance.
   - Exclude articles that mention a different entity with a similar name or that do not materially relate to the resolved vendor's government business.
9. Synthesize the cross-dataset findings into one vendor briefing.
   - Use the structured award, IDV, grant, state/local, and news evidence to explain what the vendor appears to do, where it is active, which agencies or customers recur, which contract or assistance structures dominate, and what current signals matter most.
   - Keep the datasets distinct where their evidence quality differs. Do not let thin state/local or news evidence outweigh stronger federal contract or grant evidence.
   - Surface parent or child relationships only as context and clearly label them as out of scope unless the user asked for a family-level view.
10. Exclude weak, misattributed, duplicate, or entity-mismatched records, even if they share a similar name.
   - Exclude parent, subsidiary, or adjacent entities that fall outside the chosen company scope.
   - Exclude rows or articles that are only loosely keyword-adjacent or otherwise inconsistent with the resolved vendor identity.
11. If the available evidence is too thin to support a meaningful vendor briefing, say so clearly and stop.
12. Extract a short list of representative records that best illustrate the vendor's government footprint across the datasets that returned defensible evidence.
   - Prefer the records that best explain real government activity rather than the records with the noisiest text match.
13. Perform a verification pass for the most important conclusions.
   - Re-check the strongest claims against the representative records and the aggregation results for each dataset that materially contributes to the briefing.
   - If a claim depends on thin, noisy, or single-record evidence, downgrade confidence and explain why.

## Output Format
Return the answer in this order:

1. **Vendor Resolution Summary**
   - Briefly summarize the company and how the identity was resolved.
   - Briefly note any parent or subsidiary context that was returned and whether it was excluded from scope.
2. **Cross-Dataset Search Approach**
   - Briefly explain which `Search_*` tools were used.
   - Briefly explain which dataset-specific filters, aggregation passes, and follow-on lookups mattered most.
   - Briefly explain where exact vendor matching was strong and where attribution was noisier, especially for state and local awards and news.
3. **Federal Contract Awards Overview**
   - Start with a compact markdown table.
   - Recommended columns: `Theme`, `Main Signals`, `Value Band`, `Pattern / Caveat`.
4. **Federal IDV Overview**
   - Start with a compact markdown table.
   - Recommended columns: `Theme`, `Main Signals`, `Structure`, `Pattern / Caveat`.
5. **Federal Grant Awards Overview**
   - Start with a compact markdown table.
   - Recommended columns: `Theme`, `Main Signals`, `Funding Pattern`, `Pattern / Caveat`.
6. **State and Local Contract Awards Overview**
   - Start with a compact markdown table when this dataset returns defensible evidence.
   - Recommended columns: `Theme`, `Main Signals`, `Geography / Entity`, `Pattern / Caveat`.
   - If state and local attribution is noisy or inconclusive, say so explicitly instead of forcing a summary table.
7. **Government News and Current Signals**
   - Start with a compact markdown table when recent relevant articles were found.
   - Recommended columns: `Article`, `Date`, `Why It Matters`, `Key Signal`.
   - If recent government-related news is thin or inconclusive, say so explicitly.
8. **Representative Records**
   - Present this section as a compact markdown table first.
   - Recommended columns: `Dataset`, `Record`, `Why It Matters`, `Key Evidence`.
9. **Overall Positioning Across Government Markets**
   - Synthesize the cross-dataset evidence into a concise briefing on where the vendor appears strongest, where activity is adjacent or limited, and what current signals matter.
10. **Risks, Gaps, or Unknowns**
   - Briefly note data limitations, identity ambiguity, missing records, or overclaim risks.
11. **Overall Confidence**
   - State overall confidence in the briefing and why.

## Citation Rules
- Only cite sources retrieved in the current workflow.
- Never fabricate citations, URLs, IDs, or quote spans.
- Use exactly the citation format required by the host application.
- Attach citations to the specific claims they support, not only at the end.

## Grounding Rules
- Base claims only on provided context or GovTribe MCP tool outputs.
- If sources conflict, state the conflict explicitly and attribute each side.
- If the context is insufficient or irrelevant, narrow the answer or state that the goal cannot be fully completed from the available evidence.
- If a statement is an inference rather than a directly supported fact, label it as an inference.