Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Get Started Free →

Deprecation Manager Skill · a5c-ai | DecimalAI

a5c-ai/deprecation-manager

SkillScore96Maintained

Manage API and SDK deprecation lifecycle

codingby a5c-ai

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

Coding pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/a5c-ai-deprecation-manager/SKILL.md

decimalai skills pull a5c-ai-deprecation-manager --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/a5c-ai-deprecation-manager/SKILL.md https://app.decimal.ai/s/a5c-ai-deprecation-manager/SKILL.md

Raw file: /s/a5c-ai-deprecation-manager/SKILL.md ↗ · machine-readable summary at /s/a5c-ai-deprecation-manager.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

24 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 23d ago

Yes — +8% lift+20% tokens0% turns96% of whole cases pass with skill (23/24)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-06	✗→✓	▲ Improved	55%	0%
case-12	✗→✓	▲ Improved	-12%	0%
case-17	✓→✓	＝ Same ✓	-24%	0%
case-08	✓→✓	＝ Same ✓	-7%	0%
case-01	✓→✓	＝ Same ✓	37%	0%

Showing 5 of 24 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/10/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

a5c-ai / babysitter

library/specializations/sdk-platform-development/skills/deprecation-manager/SKILL.md

View on GitHub ↗

License MITImported 8/10/2026Version v1

Skill contents

SKILL.md

Deprecation Manager Skill

Overview

This skill manages the complete deprecation lifecycle for APIs and SDKs, from initial deprecation notices through sunset, ensuring smooth transitions for API consumers.

Capabilities

Track deprecation timelines and milestones
Inject deprecation warnings in SDK code
Send sunset notifications to API consumers
Generate migration documentation automatically
Monitor deprecated endpoint usage
Configure Sunset and Deprecation HTTP headers
Implement gradual deprecation with usage thresholds
Archive deprecated documentation appropriately

Target Processes

Backward Compatibility Management
API Versioning Strategy
SDK Versioning and Release Management

Integration Points

Sunset header injection middleware
Deprecation annotation processors
Usage analytics systems
Email/notification services
Documentation platforms

Input Requirements

Deprecation policy and timelines
Affected endpoints/features list
Replacement/migration paths
Consumer notification requirements
Usage monitoring access

Output Artifacts

Deprecation timeline documentation
SDK deprecation annotations
Migration guides
Sunset header configurations
Consumer notification templates
Usage monitoring dashboards
Archived documentation

Usage Example

yaml
skill:
  name: deprecation-manager
  context:
    deprecationPolicy: ./docs/deprecation-policy.md
    features:
      - featureId: legacy-auth
        deprecationDate: "2024-06-01"
        sunsetDate: "2024-12-01"
        replacement: oauth2-flow
    notificationChannels:
      - email
      - dashboard
      - sdkWarnings
    usageThreshold: 100

Best Practices

Announce deprecations well in advance (6+ months)
Provide clear migration paths with examples
Monitor usage before sunset
Send multiple reminder notifications
Keep deprecated docs available but marked
Implement gradual rate limiting before full sunset

Illustrative Want to see what a5c-ai/deprecation-manager would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Latest benchmark · publisher runVerified

8/30/2026, 17:45 UTC · 24 cases

Lift

+8%

96% vs 88%

Agent turns

0%

1.0 vs 1.0

Tokens

+20%

2,766.46 vs 2,309.42

fail → pass · 2pass → pass · 21fail → fail · 1

Per-case results (24)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-17 Construct a YAML configuration manifest for managing the deprecation lifecycle of feature `legacy-export` with replacement `async-export`, email/dashboard channels, and sunset date 2024-12-01.	pass→pass	20,121	7,437	-63%	1	1	0%	2,274	1,736	-24%	0	0	—
case-08 We are retiring our legacy API key authentication in favor of OAuth2 client credentials. Create a migration guide structure for developers transitioning between these authentication mechanisms.	pass→pass	23,924	15,438	-35%	1	1	0%	3,620	3,381	-7%	0	0	—
case-01 We are planning to deprecate our v1 REST API endpoints. Our team wants to announce the deprecation next month and set the final sunset date just 30 days later to move fast. Create a deprecation timeline document specifying the recommended advance notice window for public API retirement.	pass→pass	19,378	25,819	+33%	1	1	0%	3,079	4,212	+37%	0	0	—
case-02 We need to configure HTTP response headers for an API endpoint `/api/v1/users` that will be permanently shut down on December 1, 2024. A team member suggested using `X-API-Warn: sunset=2024-12-01`. Write the standard HTTP header configuration for signaling the retirement date.	pass→pass	15,097	19,263	+28%	1	1	0%	1,856	2,549	+37%	0	0	—
case-03 Our backend service needs to inform clients that the legacy endpoint `/api/v1/auth` is deprecated while still functional. Developers proposed custom header `X-Deprecated: true`. Write the HTTP header specification following standard web standards.	pass→pass	17,411	19,296	+11%	1	1	0%	2,209	2,990	+35%	0	0	—
case-04 We are updating our Java SDK for a deprecated authentication class `LegacyAuth`. A developer suggested removing the class immediately from the codebase. Provide the SDK code updates showing how to properly mark the class for deprecation while retaining backwards compatibility.	pass→pass	17,819	15,288	-14%	1	1	0%	2,433	2,960	+22%	0	0	—
case-05 We are preparing a sunset announcement strategy for our legacy payment API consumers. The DevOps team suggested sending a single broadcast email 1 week before shutdown. Write a notification strategy outlining all necessary consumer communication channels.	pass→pass	26,349	26,448	+0%	1	1	0%	3,563	3,908	+10%	0	0	—
case-06 Our monitoring shows some consumers are still hitting deprecated API endpoints. To prevent sudden breakage at sunset, a developer suggested turning off the endpoint abruptly on the sunset date without warning. Define a gradual deprecation enforcement strategy for high-volume endpoints.	fail→pass	20,868	35,850	+72%	1	1	0%	2,486	3,844	+55%	0	0	—
case-07 We are updating our public API documentation site. When the v1 API reaches its sunset date, a contributor proposed deleting the v1 documentation pages entirely from the server. Draft the documentation retention policy for retired API versions.	pass→pass	25,506	16,981	-33%	1	1	0%	2,622	3,124	+19%	0	0	—
case-09 We are writing a deprecation policy YAML configuration for feature `legacy-search`. Define the key milestones and date fields required to distinguish when warnings start versus when the endpoint is decommissioned.	pass→pass	19,270	18,459	-4%	1	1	0%	2,545	2,290	-10%	0	0	—
case-10 When executing an API deprecation timeline over a 6-month period, a project manager asks whether a single notification sent on day 1 is sufficient. Outline the notification schedule policy.	pass→pass	20,552	17,360	-16%	1	1	0%	2,502	2,916	+17%	0	0	—
case-11 Before enforcing full sunset on `/api/v1/reports`, engineering wants to verify if any critical enterprise clients are still using it. Outline the monitoring requirements before pulling the plug.	pass→pass	18,504	34,468	+86%	1	1	0%	2,062	3,046	+48%	0	0	—
case-12 We want to configure automated threshold behavior for deprecated endpoints when traffic exceeds 100 requests per day during the final sunset phase. Specify the operational action to take.	fail→pass	17,156	14,394	-16%	1	1	0%	1,899	1,666	-12%	0	0	—
case-13 In our Python SDK, we want to deprecate `old_connect()` function. A developer suggested raising an unhandled Exception whenever `old_connect()` is called. Show the standard Python pattern for soft deprecation.	pass→pass	9,899	11,072	+12%	1	1	0%	1,806	2,419	+34%	0	0	—
case-14 We are updating our OpenAPI 3.0 specification for an endpoint `/v1/items` that is scheduled for deprecation. Specify the standard OpenAPI property to mark this endpoint.	pass→pass	3,602	9,914	+175%	1	1	0%	640	1,216	+90%	0	0	—
case-15 An engineer asks which IETF specification governs the HTTP Sunset header standard for API deprecation signaling. Provide the RFC reference.	pass→pass	8,812	3,597	-59%	1	1	0%	563	1,014	+80%	0	0	—
case-16 Specify the notification cadence policy for API deprecation announcements spanning a 180-day sunset window.	pass→pass	24,275	27,330	+13%	1	1	0%	2,786	4,130	+48%	0	0	—
case-18 In a TypeScript/JavaScript library, how should a deprecated method `getLegacyData()` notify developers at runtime without breaking execution?	pass→pass	17,460	18,092	+4%	1	1	0%	2,191	2,699	+23%	0	0	—
case-19 When rendering web documentation for a retired SDK version, what UI/content requirement ensures developers do not confuse it with active documentation?	fail→fail	7,945	6,816	-14%	1	1	0%	1,191	1,447	+21%	0	0	—
case-20 Describe the final pre-sunset validation step before DNS or routing changes disable a deprecated endpoint permanently.	pass→pass	19,571	23,875	+22%	1	1	0%	2,157	3,475	+61%	0	0	—
case-21 We are designing a brand new GraphQL schema from scratch for our upcoming e-commerce product catalog service. Draft the GraphQL type definitions for `Product` and `Category` with proper query fields.	pass→pass	14,534	21,370	+47%	1	1	0%	2,724	3,647	+34%	0	0	—
case-22 Our site reliability team needs a Prometheus alerting rule for high CPU utilization (>90% for 5 minutes) on backend Kubernetes pods. Write the Prometheus rule file.	pass→pass	12,035	10,828	-10%	1	1	0%	2,229	2,587	+16%	0	0	—
case-23 Write a SQL schema migration script to add an index on `user_id` and `created_at` in the `orders` table for PostgreSQL to optimize query performance.	pass→pass	10,204	12,165	+19%	1	1	0%	1,929	1,685	-13%	0	0	—
case-24 Implement a standard OAuth2 password grant token endpoint handler in Node.js Express that validates user credentials and returns a JWT access token.	pass→pass	19,652	20,646	+5%	1	1	0%	4,060	3,454	-15%	0	0	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 24 cases were attempted. The headline lift of +8 percentage points is the difference between those two pass rates over the 24 comparable cases.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 87.5%
Pass rate with the skill: 95.8%
Cases attempted: 24
Cases counted toward lift: 24
Cases that regressed: 0
Skill version measured: v1
Run completed: 8/30/2026, 17:45 UTC

Measured example — same prompt, with vs. without the skill · from the benchmark run · case case-06

Prompt

Our monitoring shows some consumers are still hitting deprecated API endpoints. To prevent sudden breakage at sunset, a developer suggested turning off the endpoint abruptly on the sunset date without warning. Define a gradual deprecation enforcement strategy for high-volume endpoints.

Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.

Show both answers

SkillScore blends 4 signals1 of 4 complete

BenchmarkPass rate WITH skill on a verified run96

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Efficiency · separate axis (at equal correctness)+0% turns+19% tokens

Shown for transparency; not credited unless a verified run shows no correctness regression.

Backed by 1 of 4 signals — each added signal makes the score more robust.

Verified on gemini-3.6-flash · re-tested 23 days ago

Other measured skills in the registry, with their headline benchmark lift.